Comprehensive research synthesis · Version 1.1 · 17 July 2026

Build communication as durable enterprise infrastructure.

A mobile-first, evidence-linked reference for academic research, technical documentation, architecture explanations, personas, storytelling, persuasion, business value, temporal trade-offs, presentations, marketing, newsletters, LLM retrieval, and future-agent research continuity.

Explore findings →Use a playbook

Human-readableMachine-readableEvidence-linkedPersona-awareArgument-awareVersionableRed-team ready

Directed and owned by Jesse GraupmannResearch, synthesis, implementation, and validation assisted by OpenAI ChatGPT · GPT-5.6 Thinking. Source creators remain independently attributed.

Operating model

The research compiler that turns durable knowledge into audience-specific communication.

The core conclusion

The durable solution is not a better universal template. It is a governed communication system that stores research and reasoning independently from the articles, HTML explainers, presentations, specifications, newsletters, and LLM answers generated from it.

Evidence and sources

Canonical knowledge and arguments

Audience and impact contract

Channel-specific output

The compiler analogy

Raw inputs: Papers, standards, operational data, interviews, company practices, and source annotations.
Abstract syntax tree: Claims, concepts, evidence, warrants, decisions, examples, procedures, personas, and trade-offs.
Rendering profiles: Audience, situation, intent, evidence threshold, medium, depth, tone, and CTA.
Compiled outputs: Article, specification, architecture view, deck, executive memo, newsletter, landing page, or retrieval chunk.
Source maps: Citations, provenance, exact locators, versions, and epistemic change history.

Primary quality contract

Correctness: facts and technical statements are accurate.
Relevance: the view includes what the declared reader needs.
Traceability: claims and decisions link to evidence and reasoning.
Findability: humans and machines can locate the appropriate unit.
Actionability: the reader can complete the intended decision or task.
Currency: owner, status, version, and verification date are explicit.
Accessibility: structure and meaning survive alternative modalities.
Governability: changes, challenges, supersession, and retirement are controlled.

Research domains

Reader intent and information architecture

Orient, teach, help act, support decisions, enable verification, operate, persuade, and retrieve.

View findings

Academic and research structures

IMRaD, CARS, literature reviews, claim–evidence–analysis, method, results, and limitations.

View findings

Evidence and argumentation

Claims, evidence, warrants, backing, qualifiers, counterarguments, and confidence.

View findings

Technical documentation and specifications

Tutorials, how-to guides, reference, explanation, runbooks, specifications, ADRs, and lifecycle content.

View findings

Architecture communication

Stakeholder concerns, viewpoints, diagrams, narrative explanations, boundaries, and operational implications.

View findings

Storytelling and presentations

Controlled change, data narrative, assertion–evidence slides, companion documents, and cognitive load.

View findings

Marketing, newsletters, and conversion

Category, relevance, value, proof, objections, specific CTA, recurring utility, and premium differentiation.

View findings

Personas and audience impact

Populations, segments, archetypes, situations, impact contracts, validation, ethics, and lifecycle.

View findings

Persuasion and persuasive essays

Logos, ethos, pathos, elaboration, reactance, inoculation, framing, uncertainty, and argument structures.

View findings

Business value and temporal trade-offs

Net value, beneficiary and cost bearer, reversibility, option value, debt, lock-in, and time horizons.

View findings

LLM retrieval and AEO

Semantic chunks, local context, reranking, validation, authoritative sources, and model-independent adapters.

View findings

Research continuity and future agents

Atomic research objects, provenance, source preservation, challenge records, evaluation suites, and open packaging.

View findings

Enterprise governance and measurement

Owners, status, review triggers, permitted use, prohibited use, success measures, and retirement.

View findings

Foundational principles

The rules that remain stable across document types, channels, and technologies.

P01

Intent before format

Choose the reader task and decision before choosing a document, page, diagram, or deck.

DIATAXIS GOOGLE-HEADINGS

P02

Evidence before authority

Make source quality, applicability, uncertainty, and reasoning inspectable. Expertise is demonstrated through traceability.

UNC-EVIDENCE SOURCE-CREDIBILITY

P03

Preserve depth, publish layers

Keep full evidence and provenance in the canonical layer. Publish progressively reduced views for each audience and channel.

REDHAT-MODULAR MAYER-MULTIMEDIA

P04

Build knowledge once; render many ways

Separate claims, concepts, evidence, decisions, examples, and procedures from individual output layouts.

REDHAT-MODULAR FAIR

P05

Structure is meaning

Semantic headings, relationships, labels, and content boundaries serve humans, assistive technology, search, and retrieval systems.

WCAG-INFO-REL GOOGLE-HEADINGS

P06

Explain why before how

Connect purpose, problem, evidence, causality, priority, choice, and accepted trade-offs before implementation detail.

SWALES-CARS PURDUE-TOULMIN

P07

Trade-offs over advocacy

Show what improves, what worsens, who benefits, who pays, and what future options are preserved or constrained.

REAL-OPTIONS SEI-TECH-DEBT

P08

Personas are bounded models

Use personas only for declared decisions and contexts. Prefer behavioral evidence over invented biography.

MICROSOFT-PERSONAS CHAPMAN-MILHAM

P09

Make reasoning inspectable

Store claims, grounds, warrants, qualifiers, assumptions, rebuttals, alternatives, and decision criteria.

PURDUE-TOULMIN AIF

P10

Define measurable audience impact

State what the audience should know, decide, do, verify, or explain after using an artifact.

WHO-ACTIONABLE IBM-RESEARCH-PLANNING

P11

Accessibility and ethics are architectural

Do not reserve semantic structure, inclusive participation, privacy, or persuasion integrity for final review.

WCAG-INFO-REL GOVUK-RESEARCH-PRIVACY

P12

Version knowledge; do not overwrite it

Supersede claims and recommendations with an epistemic change log. Preserve rejected challenges and prior states.

W3C-PROV FAIR

Communication atlas

Choose a genre by the reader contract, not by organizational habit.

Each form optimizes for a different type of work. The canonical facts may be shared, but the hierarchy, level of detail, evidence presentation, and call to action must change.

Form	Use it for	Core structure	Do not use it as
Research note	Capture one observation, source, or question.	Question → observation → source → interpretation → follow-up	Not a final conclusion.
Annotated bibliography	Record what a source contributes and how reliable it is.	Citation → summary → quality → useful claims → limitations	Do not treat separate annotations as synthesis.
Literature review	Map existing knowledge, debates, methods, and gaps.	Scope → method → themes → agreement → conflict → gaps → implications	Do not organize only by source.
Research paper	Contribute an original argument, method, or result.	Introduction → method → results → discussion → limitations	Not an operational procedure.
Experiment or benchmark report	Make a test reproducible and its result bounded.	Hypothesis → environment → method → metrics → results → interpretation	Do not generalize beyond workload and environment.
Case study	Explain an intervention and outcome in context.	Context → problem → intervention → outcome → limits → transferable lesson	One case is not universal proof.
White paper	Educate and establish a defensible direction.	Problem → evidence → approach → value → implementation → next action	Do not present advocacy as independent academic evidence.
Executive brief	Enable a rapid consequential decision.	Decision → significance → evidence → options → recommendation → risk	Do not make the decision request implicit.
Narrative memo	Develop shared context and reasoning.	Situation → tension → evidence → alternatives → recommendation → consequences	Not ideal for primarily visual demonstrations.
Business case	Justify investment or prioritization.	Problem → strategic fit → options → net value → risk → recommendation	Do not count transferred costs as savings.
Proposal	Request permission, funding, or commitment.	Need → response → scope → deliverables → resources → risk → acceptance	Not a status report.
RFC	Collect broad review before a meaningful change.	Summary → motivation → design → alternatives → compatibility → rollout → open questions	Avoid for trivial local choices.
ADR	Preserve one significant architecture decision.	Context → decision → status → consequences → revisit conditions	Not a substitute for the complete system design.
PRD	Align on user and product outcomes.	Problem → goals → requirements → non-goals → measures → constraints	Do not bury technical implementation decisions inside product language.
Functional specification	Describe expected behavior.	Actors → scenarios → behavior → rules → errors → acceptance	Does not replace an implementation design.
Technical specification	Provide an implementation and operational blueprint.	Context → requirements → architecture → interfaces → security → operations → rollout	Premature when the solution remains exploratory.
Tutorial	Help a learner acquire competence.	Outcome → prerequisites → guided sequence → checkpoints → result → next step	Do not double as exhaustive reference.
How-to guide	Help a competent reader accomplish a task.	Goal → prerequisites → steps → expected result → recovery	Do not teach the whole conceptual domain.
Reference	Support exact lookup.	Definition → syntax → parameters → constraints → examples → errors	Do not force a learning journey.
Explanation	Build a mental model.	Definition → relationships → mechanism → examples → boundaries	Do not disguise procedures as theory.
Runbook	Support reliable operation and recovery.	Trigger → diagnosis → action → validation → escalation → rollback	Do not bury it in architecture prose.
Postmortem	Learn from an incident and improve controls.	Impact → timeline → causes → contributing conditions → actions → verification	Avoid blame and retrospective certainty.
Thought-leadership article	Establish a defensible point of view.	Thesis → context → evidence → synthesis → implications	Do not use novelty of tone as evidence.
Landing page	Move a qualified reader toward a bounded action.	Identity → relevance → outcome → mechanism → proof → constraints → action	Do not ask for conversion before understanding.
Newsletter	Deliver recurring signal and utility.	Issue thesis → findings → interpretation → links → action	Do not reproduce the entire research report.
Presentation	Guide a time-bound spoken argument.	Assertion → evidence → transition → decision	A deck is not the durable source of truth.
Policy	State mandatory organizational requirements and rationale.	Purpose → scope → requirements → responsibilities → exceptions → enforcement	Do not use should when must is intended.
Guideline	Provide context-sensitive recommended practice.	Context → recommendation → rationale → exceptions → examples	Do not make optional guidance appear mandatory.

Argument and research structures

Structure	Best use	Sequence	Benefit
IMRaD	Experimental or systematic research	Introduction → Methods → Results → Discussion	Separates what was done, observed, and inferred. GMU-IMRAD
CARS	Research and proposal introductions	Establish territory → establish niche → occupy niche	Explains why the contribution is necessary. SWALES-CARS
Claim–evidence–analysis	Research paragraphs and enterprise recommendations	Assertion → evidence → interpretation → implication → action	Prevents citation dumping without reasoning. HARVARD-ORGANIZING UNC-ARGUMENT
Toulmin	Conditional practical recommendations	Claim → grounds → warrant → backing → qualifier → rebuttal	Exposes hidden reasoning and exceptions. PURDUE-TOULMIN
Classical argument	Strategic advocacy and major presentations	Relevance → context → position → proof → refutation → action	Creates a broad persuasive arc. PURDUE-CLASSICAL
Rogerian argument	Legitimate stakeholder conflict	Neutral issue → opposing view → valid conditions → own view → shared goals → resolution	Finds common ground without erasing trade-offs. PURDUE-ROGERIAN
Policy memo	Executive decision	Decision → evidence → options → comparison → recommendation → implementation	Optimizes for a specific decision-maker. POLICY-MEMO
Technical story	Engineering articles and transformation narratives	Context → friction → failed obvious answer → insight → decision → implementation → outcome → limits	Uses change to explain causality and learning. NARRATIVE-TRANSPORT
Data story	Analytical communication	Question → baseline → contrast → explanation → consequence → action	Moves from measurement to decision. DATA-STORY-REVIEW
Assertion–evidence	Presentations	Complete-sentence assertion + supporting visual evidence	Reduces topic headings and bullet walls. ASSERTION-EVIDENCE
Enterprise why chain	Recommendations	Purpose → problem → evidence → causality → priority → choice → trade-off → how	Explains why before implementation. PURDUE-TOULMIN SWALES-CARS

Genre selection rule

A document that tries to orient, teach, guide action, provide exact reference, justify a decision, market a product, and preserve research provenance in one visible hierarchy will usually underperform every one of those jobs. Share canonical objects; separate the views.

Academic and enterprise research

How to collect, evaluate, synthesize, argue, and preserve evidence.

Evidence and authority model

Subject-matter expertise is demonstrated by making the boundary between known, inferred, recommended, and uncertain material visible.

Class	Appropriate meaning
Standard	Normative requirement from a recognized standards body.
Systematic synthesis	Review or meta-analysis across multiple studies.
Primary research	Original study, experiment, benchmark, or empirical analysis.
University guidance	Research- and discipline-informed writing or method guidance.
Company practice	First-party documentation of a named organization’s implementation.
Framework	Reusable professional model that may not be universally empirical.
Our synthesis	A conclusion derived across several sources and enterprise constraints.
Hypothesis	A proposition requiring validation.
Subjective judgment	An editorial, visual, or strategic preference.

Atomic argument unit

Assertion

Evidence

Interpretation

Enterprise implication

For consequential recommendations, extend the unit with a warrant, backing, qualifier, counterargument, limitations, and a recommendation.

Boundary: a cited fact is not automatically evidence for the recommendation. The causal or decision rule connecting them must be explicit.

Source hierarchy

Normative standards and specifications.
Systematic reviews and meta-analyses.
Original peer-reviewed research.
University and professional research centers.
First-party engineering and product practices.
Government guidance.
Established professional frameworks.
Secondary commentary and anecdotal examples.

The order is contextual. A company engineering article is the primary source for that company’s implementation, but not necessarily strong evidence for a universal behavioral claim.

Research stages

1. Frame

Research objective, intended decision, audiences, time horizon, scope, exclusions, terminology, evidence bar, and contested questions.

2. Collect and assess

Search broadly; evaluate authority, method, relevance, recency, independence, reproducibility, applicability, and conflicts.

3. Extract

Capture claim supported, exact evidence, method, population, limitations, significance, contradictory findings, and exact source locator.

4. Synthesize

Group established principles, convergence, disagreement, context-dependent practice, emerging evidence, gaps, and enterprise implications.

5. Validate

Fact, citation, technical, adversarial, accessibility, readability, privacy, security, retrieval, and stakeholder review.

6. Publish and maintain

Compile audience-specific outputs; measure outcomes; define owner, review triggers, expiration, supersession, and archive policy.

Technical documentation and architecture

From reader journeys to specifications, decisions, diagrams, operations, and lifecycle.

Documentation system

Documentation should be treated as a product with users, journeys, requirements, quality attributes, defects, analytics, operational ownership, release cycles, and deprecation.

Learn: tutorial.
Perform: how-to guide or procedure.
Look up: reference.
Understand: explanation.
Operate: runbook and troubleshooting.
Decide: architecture overview, RFC, specification, and ADR.

DIATAXIS REDHAT-MODULAR SPOTIFY-GOLDEN-PATHS

Documentation quality attributes

CorrectnessTechnically and factually accurate.

CompletenessContains what the declared purpose requires.

FindabilityCan be located through navigation, search, and retrieval.

ActionabilitySupports the intended task or decision.

TraceabilityLinks requirements, decisions, evidence, and implementation.

CurrencyOwner, verification date, version, and replacement status are known.

OperabilityFailure, diagnosis, recovery, escalation, and support are documented.

Enterprise specification minimum

Layer	Required coverage
Control	ID, owner, status, reviewers, version, classification, last verified.
Context	Current state, user and business need, trigger, goals, non-goals, constraints, assumptions.
Architecture	Components, data lifecycle, interfaces, identity, authorization, trust boundaries, dependencies.
Quality	Security, privacy, compliance, accessibility, capacity, performance, resilience, disaster recovery.
Lifecycle	Testing, migration, rollout, rollback, observability, support, maintenance, ownership.
Decision	Alternatives, trade-offs, consequences, open questions, success measures, references.

Architecture-view contract

A diagram is a view, not the architecture. Every diagram declares intended audience, stakeholder concern, scope, abstraction level, time perspective, included and omitted elements, notation, source of truth, owner, and last verification.

View	Question answered
Landscape	What systems and domains exist?
System context	Who and what interacts with the system?
Container	What deployable or executable units compose it?
Component	What responsibilities exist inside a unit?
Dynamic or sequence	What happens in a scenario?
Deployment	Where does software run?
Data flow and lineage	Where does information originate, transform, persist, and propagate?
Threat model	Where are trust boundaries, threats, controls, and residual risk?
Operational topology	How is the service observed, supported, scaled, and recovered?

Personas and audience impact

Research-backed audience models that guide communication without becoming persuasive fiction.

Definition

A persona is a versioned, evidence-backed representation of a decision-relevant behavioral pattern within a defined population, context, and period.

A persona is not

An average user or a real participant.
A market-size estimate or authorization role.
A job description or demographic stereotype.
A permanent truth about a population.
A replacement for direct research.
An instruction for an LLM to invent audience evidence.

Use the least fictional model

Enterprise work usually benefits from an archetype such as time-constrained service operator investigating an active production failure more than a named character with an invented age, photograph, family, and hobbies.

NNG-PERSONAS-ARCHETYPES PERSONA-STEREOTYPE PERSONA-QUANT-TEST

Audience constellation

Audience role	Communication concern
Primary consumer	The task that determines the artifact’s principal organization.
Decision maker	Evidence, consequences, alternatives, and decision criteria.
Implementer	Contracts, examples, failure behavior, acceptance, and migration.
Operator	Telemetry, ownership, failure, recovery, escalation, and support.
Reviewer or gatekeeper	Correctness, security, risk, compliance, funding, or procurement.
Affected party	Consequences experienced despite not directly using the artifact or system.
Future maintainer	Historical rationale, supersession, and reconstruction of context.
Adversarial actor	Ambiguity, weak controls, and exploitable assumptions.

Impact objective formula

After consuming [artifact], members matching [persona and situation] should be able to [observable outcome] under [conditions] with [quality threshold], as measured by [indicator], without [dangerous interpretation or behavior].

Persona research safeguards

Define the decision domain and prohibited uses before collecting data.
Discover behavioral dimensions before naming personas.
Combine qualitative explanation with quantitative testing or prevalence estimation where feasible.
Include contrast cases, negative evidence, holdout validation, and people who fit multiple or no personas.
Separate restricted participant evidence from de-identified observations and aggregated persona output.
Label attributes as observed, reported, measured, inferred, hypothesized, illustrative, or deprecated.
Version and revise personas when research, product, organization, regulation, or behavior changes.

Synthetic-persona policy: model-generated personas may help brainstorm questions or objections. They are not interviews, survey respondents, market prevalence, accessibility validation, or evidence of likely production behavior.

Persuasion, business outcomes, and trade-offs

Design recommendations as inspectable arguments that improve informed choice.

Ethical standard

Enterprise persuasion should improve the quality of informed choice rather than merely increase agreement with the author.

Logos

Evidence, causal reasoning, financial analysis, architecture logic, trade-offs, assumptions, alternatives, and measures.

Ethos

Accuracy, transparent uncertainty, fair alternatives, accountability, competence, independence, and verifiability.

Pathos

Make customer harm, operator burden, risk, lost opportunity, confidence, and urgency tangible without exceeding the evidence.

Inspectability chain

Problem and significance

Claim and evidence

Warrant and alternatives

Recommendation and exit

Persuasive safeguards

Layer rapid orientation over full substantive analysis.
Assess source credibility for each claim and decision context.
Avoid controlling language that suppresses legitimate autonomy.
Prebunk credible objections and identify where they are valid.
Reconnect stories to base rates and aggregate evidence.
Test gain, loss, cost, status-quo, and downside frames.
Quantify uncertainty and identify its drivers.

Complete why chain

Purpose: what objective matters?
Problem: what is insufficient?
Evidence: how do we know?
Causality: why should the intervention work?
Priority: why now?
Choice: why this option?
Trade-off: why are the disadvantages acceptable?
How: architecture, ownership, controls, migration, and operation.

Business value dimensions

Financial

Revenue, margin, avoided cost, cash flow, capital efficiency

Customer

Retention, satisfaction, availability, speed, trust

Operational

Throughput, reliability, recovery, support burden

Risk

Probability, impact, exposure, compliance, residual risk

Strategic

Differentiation, time to market, option value, adaptability

Technical

Evolvability, interoperability, resilience, maintainability

Workforce

Cognitive load, onboarding, productivity, retention

Learning

Evidence, reusable capability, institutional knowledge

Governance

Visibility, consistency, ownership, auditability

Ecosystem

Partner integration, reuse, portability, standards alignment

Net-value contract

Net value = benefits − direct costs − operating costs − transition costs − opportunity costs − risk exposure − future constraints.

Every value claim identifies outcome, beneficiary, mechanism, time horizon, measure, baseline, uncertainty, and cost bearer.

Time-expanded trade-off model

Horizon	Questions
Immediate	Delivery, disruption, migration effort, decision overhead, and early failure.
Near term	Adoption, training, stability, initial value, and support burden.
Medium term	Operating cost, scaling, debt, organizational dependency, and measurement.
Long term	Strategic flexibility, architecture evolution, lock-in, and market or regulatory change.
Exit	Replacement, data portability, retained contracts, decommissioning, and restoration of the prior state.

Trade-off dimensions

Outcome · evidence · beneficiary · cost bearer · time to value · duration · direct and opportunity cost · transition and operating cost · probability and impact · reversibility · option value · path dependency · technical debt · scalability · security · governance · distribution · confidence.

Reusable output playbooks

Atomic structures that can be mixed and matched without changing canonical facts.

These are assembly patterns. They specify hierarchy and decision purpose. They do not replace research, source assessment, editorial judgment, accessibility testing, or technical validation.

Reader-intent router

Select the correct communication form before authoring.

OrientLandscape, glossary, system context, executive summary

LearnTutorial with guided sequence and checkpoints

ActHow-to, procedure, checklist, or runbook

DecideMemo, business case, RFC, or options analysis

VerifyReference, specification, standard, or source registry

UnderstandExplanation, architecture narrative, or literature synthesis

PersuadeArgument, proposal, landing page, or presentation

RetrieveAtomic answer unit with evidence and local context

Literature review

Synthesize a field without becoming a list of sources.

1Define scope, questions, terminology, and selection method

2Map theories, methods, populations, and source quality

3Organize themes, convergence, disagreement, and gaps

4Separate evidence from enterprise transfer inference

5Conclude with implications, limitations, and research agenda

IMRaD research report

Preserve methodological inspectability.

IntroductionProblem, prior knowledge, gap, and research question

MethodsPopulation, environment, instruments, variables, and analysis

ResultsObservations without advocacy

DiscussionInterpretation, limitations, transferability, and recommendation

Enterprise technical specification

Connect design to implementation, operations, and outcomes.

ControlID, owner, status, reviewers, version, classification

ContextCurrent state, trigger, goals, non-goals, constraints

DesignArchitecture, components, data, interfaces, identity

QualitySecurity, privacy, accessibility, performance, resilience

LifecycleTesting, migration, rollout, rollback, operations, ownership

DecisionAlternatives, trade-offs, open questions, measures, references

Architecture view contract

Make every diagram answer a declared concern.

AudienceWho reads this view?

ConcernWhat question does it answer?

ScopeWhat is included, excluded, and abstracted?

NotationLegend, boundaries, direction, and time perspective

NarrativeReading order, primary flow, decisions, failures, operations

LineageSource definitions, ADRs, owner, and last verified

Persona Stack

Prevent a static profile from substituting for context.

PopulationWho may be affected?

SegmentWhich measurable pattern matters?

PersonaWhat recurring behavioral decision pattern exists?

SituationWhat is happening now?

BaselineWhat is known, believed, and currently done?

ImpactWhat should change after communication?

ConstraintsTime, device, environment, authority, accessibility

Persona foundation package

Turn a persona into governed research infrastructure.

Operational cardConcise daily reference

Foundation documentFull research basis, scope, synthesis, and limitations

Evidence matrixAttribute-level observed, measured, inferred, and hypothesized evidence

Scenario libraryTasks and situations where the model applies

Impact contractsDesired knowledge, decision, behavior, and guardrails

LifecycleVersion, validation, drift signals, and retirement

Persona Impact Contract

Define a measurable communication outcome.

BaselineKnowledge, attitude, behavior, authority, and barriers

Desired cognitionKnow, distinguish, remember, and explain

Desired decisionCompare, approve, reject, prioritize, or escalate

Desired behaviorStart, stop, complete, adopt, configure, or share

Evidence thresholdProof required for trust and action

GuardrailDangerous interpretation or action to prevent

MeasurementIndicator, threshold, time horizon, and dependencies

Toulmin argument

Make a recommendation inspectable.

ClaimWhat exactly is recommended?

GroundsWhat evidence supports it?

WarrantWhy does the evidence imply the claim?

BackingWhy should the warrant be accepted?

QualifierWhere and with what confidence does it apply?

RebuttalWhen is the claim invalid or incomplete?

Enterprise why chain

Explain why before how.

PurposeWhat human or organizational objective matters?

ProblemWhat is insufficient today?

EvidenceHow do we know?

CausalityWhy should the intervention work?

PriorityWhy now?

ChoiceWhy this option?

Trade-offWhy are disadvantages acceptable?

HowArchitecture, ownership, controls, migration, and operation

Decision paper

Enable a defensible enterprise choice.

DecisionDecision requested and recommendation

StakesWhy now and consequence of inaction

EvidenceCurrent-state findings and root causes

OptionsStatus quo, alternatives, and evaluation criteria

ValueOutcomes, beneficiary, mechanism, baseline, and uncertainty

Trade-offsShort, medium, long, exit, reversibility, and lock-in

ExecutionOwnership, milestones, controls, validation, and stop conditions

Time-expanded trade-off analysis

Prevent short-term value from hiding future cost.

ImmediateDelivery, disruption, migration, and approval

Near termAdoption, training, stability, and initial value

Medium termOperating cost, scaling, debt, and dependencies

Long termStrategic flexibility, lock-in, and architecture evolution

ExitReplacement, decommissioning, portability, and residual obligations

Assertion–evidence presentation

Build a deck around claims rather than topics.

AssertionOne complete sentence that advances the argument

EvidenceA diagram, chart, image, or bounded data display

NarrationWhat cannot be inferred from the visual alone

SourceVisible or note-level citation

TransitionWhy the next assertion follows

CompanionDurable document with methods, caveats, and detail

Research workflow

Move from question to governed publication.

FrameObjective, decision, audience, scope, terms, evidence bar

CollectStandards, reviews, primary studies, first-party cases, commentary

AssessAuthority, method, relevance, independence, applicability

ExtractClaim, evidence, method, limits, contradiction, locator

SynthesizeConvergence, disagreement, context, gaps, implications

ValidateFact, citation, technical, adversarial, accessibility, retrieval

PublishOutput manifest and channel-specific assembly

MaintainOwner, verification interval, supersession, and archive

Future Research Continuity Package

Enable reproduction, audit, red-team, and extension.

SourcesImmutable evidence, snapshots, hashes, licenses, locators

RegistrySource quality, methods, scope, conflicts, claims supported

KnowledgeAtomic claims, evidence, concepts, assumptions, limitations

ArgumentsWarrants, qualifiers, rebuttals, alternatives, decisions

AudiencesPersonas, situations, impact contracts, affected parties

ProvenanceAgents, activities, derivations, prompts, reviews

ChallengesOpen, resolved, and rejected adversarial findings

EvaluationsCitation, calibration, trade-off, persona, temporal tests

PackageREADME, schemas, RO-Crate metadata, manifest, changelog

Future-agent red-team protocol

Challenge without silently rewriting history.

IntegrityVerify sources, versions, hashes, quotations, and dates

EvidenceCheck claim scope, source support, causality, and applicability

ArgumentChallenge warrants, assumptions, qualifiers, and rebuttals

AlternativesAdd status quo, hybrid, and omitted options

Trade-offsExpose transferred cost, time horizon, reversibility, and lock-in

AudienceIdentify missing personas, affected parties, and accessibility

FreshnessCheck superseded standards, new research, and changed practice

ExpansionAppend sources, claims, challenges, and a versioned synthesis

Content compiler

Render one knowledge base into many channels.

Canonical objectsClaims, concepts, evidence, decisions, procedures, examples

Audience contractPersona, situation, baseline, authority, and desired impact

Output manifestGenre, hierarchy, included objects, evidence rules, CTA

RendererArticle, SPA, deck, memo, docs, newsletter, or retrieval chunk

ValidatorFacts, citations, terminology, accessibility, and policy

FeedbackObserved outcome updates content and audience models

Output-specific rules

Output	Lead with	Preserve elsewhere
Research article	Thesis, significance, method, thematic findings, implications.	Source registry, extraction notes, complete claim ledger.
Interactive SPA	Orientation, search, filters, progressive disclosure, stable anchors.	Embedded machine data, source details, version history, print view.
Presentation	Assertions, visual evidence, spoken sequence, decision.	Companion document, notes, appendix, citations, methods.
Executive memo	Decision, why now, recommendation, value, principal trade-off.	Technical design, calculations, evidence appendix.
Technical documentation	Reader task: learn, act, understand, look up, or recover.	Canonical concepts and contracts reused across views.
Newsletter	One signal, significance, evidence, action.	Canonical deep dive and source object.
Marketing page	Category, audience relevance, outcome, mechanism, proof, constraint, action.	Claim classification and substantiation record.
LLM response	Direct answer with local context, scope, evidence, and uncertainty.	Retrieved objects, prompt, citations, validation, and provenance.

Canonical knowledge and future-agent system

How research remains auditable, extensible, model-independent, and safe to transform.

Four continuity operations

Reproduce what the original researchers saw and did. Audit whether claims follow from evidence. Red-team errors, assumptions, omissions, and alternatives. Extend the work without erasing the prior state.

Atomic research objects

sourceExternal or internal artifact with bibliographic and quality metadata.

annotationExact passage, figure, table cell, or resource segment linked to a claim.

claimA supportable statement with scope, confidence, and evidence relationships.

evidenceObservation, result, standard, or quotation supporting or challenging a claim.

warrantReasoning that connects evidence to a claim or recommendation.

assumptionA condition currently treated as true but requiring monitoring.

conceptA reusable definition or mental model.

interpretationWhat evidence means within a declared context.

recommendationA proposed action with outcomes, trade-offs, owner, and revisit conditions.

decisionA selected option with context and consequences.

counterargumentA credible objection or competing interpretation.

limitationA boundary on validity or applicability.

procedureAn ordered action sequence with inputs, controls, and validation.

metricA measurement, baseline, threshold, and interpretation rule.

personaA bounded evidence-backed audience decision model.

situationA context in which an audience pursues a job or makes a decision.

impact-contractThe intended measurable change in audience knowledge, judgment, or action.

challengeA structured adversarial finding against a research object.

output-manifestThe rules assembling canonical objects for a specific channel and audience.

provenance-activityA human or model transformation with inputs, outputs, and review.

Research lineage

Source paper or operational evidence
    └─ supports / disputes / qualifies → Claim
          └─ connected by → Warrant and assumptions
                └─ compared through → Options and trade-offs
                      └─ generates → Recommendation or decision
                            └─ assembled for → Persona + situation + impact contract
                                  └─ renders → Article / SPA / deck / spec / newsletter / RAG unit
                                        └─ produces → Measured audience and business outcomes
                                              └─ triggers → Challenge / revision / supersession

Recommended package

enterprise-communication-research/
├── README.md
├── ro-crate-metadata.json
├── manifest.json
├── research-agenda/
├── sources/
│   ├── registry.yaml
│   ├── assessments/
│   ├── snapshots/
│   └── annotations/
├── knowledge/
│   ├── concepts/
│   ├── claims/
│   ├── evidence/
│   ├── warrants/
│   ├── assumptions/
│   ├── limitations/
│   └── recommendations/
├── audiences/
├── deliberation/
├── outputs/
├── challenges/
├── evaluations/
├── provenance/
└── schemas/

Format contract

Human canonical: Markdown.
Structured canonical: JSON or YAML validated with JSON Schema.
Semantic exchange: JSON-LD for provenance and relationships.
Tabular: CSV or Parquet.
Diagrams: Mermaid, PlantUML, Graphviz, or another text source.
Identity: stable logical IDs, semantic versions, hashes, and parent relationships.
Not canonical alone: PDF, slide deck, screenshot, proprietary whiteboard, embedding index, or chat transcript.

Machine-readable schema examples

Claim object

id: claim-content-modularity-001
object_type: claim
canonical_statement: >
  Content organized as bounded semantic modules can be assembled
  into multiple audience- and task-specific outputs.
status: verified
evidence_class: company-practice
confidence: medium
scope:
  domains: [technical-documentation, enterprise-knowledge]
  exclusions: [literary-long-form]
evidence:
  supporting:
    - source_id: REDHAT-MODULAR
      relation: supports
    - source_id: DIATAXIS
      relation: corroborates
warrant: >
  Separating reusable knowledge from a page layout permits controlled
  recombination without changing the underlying claim.
limitations:
  - Modules still require editorial assembly and connective context.
  - Excessive fragmentation can damage narrative coherence.
provenance:
  created_at: 2026-07-17
  last_verified_at: 2026-07-17
governance:
  owner: platform-content
  review_cycle_days: 180

Persona object

id: persona-platform-adopter-001
object_type: persona
version: 2.1
status: active
label: Pragmatic platform adopter
scope:
  decision_domain: [platform-adoption, technical-documentation]
  prohibited_uses: [market-sizing, access-control, performance-evaluation]
research_basis:
  methods: [interviews, architecture-review-analysis, support-analysis]
  sample:
    participants: 18
    limitations: [overrepresentation-of-early-adopters]
behavioral_model:
  primary_goal: Adopt a maintainable integration without losing control.
  primary_tension: Standardization helps, but opaque abstraction creates risk.
  decision_criteria: [reliability, observability, migration-effort, security, reversibility]
  evidence_preference: [measured-results, explicit-contracts, failure-scenarios]
communication_contract:
  baseline: {knowledge: moderate, trust: skeptical}
  desired_impact:
    understand: [ownership-boundaries, routing-and-failure-model]
    decide: [whether-to-run-pilot]
    act: [complete-migration-assessment]
  prohibited_messaging: [seamless, zero-risk, eliminates-lock-in]
validation:
  unresolved_questions: [prevalence, regulated-team-differences]

Argument object

id: argument-ai-gateway-default-001
object_type: argument
claim: Adopt a shared AI gateway as the default model-access path.
grounds:
  - Teams duplicate provider authentication, logging, retry, policy, and reporting.
warrant: >
  Capabilities repeated across services can be governed and maintained more
  consistently as a supported platform capability.
backing:
  - internal-duplication-analysis
  - incident-history
  - platform-engineering-research
qualifier: >
  Applies to standard inference workloads within the measured latency envelope.
rebuttal:
  - Unsupported provider capabilities may justify direct integration.
  - Hard real-time or legally isolated workloads may require exceptions.
assumptions:
  - The platform has funded operational ownership.
  - Emergency bypass and rollback are tested.
status: proposed

Challenge object

id: challenge-persuasive-essay-004
object_type: challenge
type: overgeneralization
target:
  claim_id: claim-narrative-persuasion-002
  version: 1.1.0
challenge: >
  The cited evidence contains consumer and health communication studies;
  direct enterprise transfer is not yet established.
basis: [scope-mismatch, missing-enterprise-validation]
severity: medium
status: open
proposed_resolution: [narrow-scope, add-enterprise-study, label-transfer-inference]
raised_by:
  agent: future-model-id
  date: 2027-08-11

Future-agent handoff

research_handoff:
  title: Enterprise Communication Research System
  version: 1.0.0
  status: active-research
  last_reviewed: 2026-07-17
  purposes: [preserve, reproduce, audit, red-team, expand]
  start_here:
    human: README.md
    agent: manifests/agent-entry.yaml
  source_of_truth:
    claims: knowledge/claims/
    sources: sources/
    provenance: provenance/
    decisions: deliberation/decisions/
  known_weaknesses:
    - limited non-Western rhetoric research
    - incomplete empirical evaluation of enterprise persona impact
    - limited longitudinal measurement of documentation outcomes
  prohibited_agent_behavior:
    - do not delete prior claims
    - do not convert inference to fact
    - do not upgrade confidence without evidence
    - do not silently repair citations
  required_agent_outputs:
    - findings
    - citations
    - contradictions
    - proposed_changes
    - confidence
    - provenance_record

Output assembly manifest

id: architecture-article-ai-gateway
object_type: output-manifest
output_type: long-form-technical-article
purpose: {primary: explain, secondary: [build-confidence, support-decision]}
audience_constellation:
  primary_persona: persona-platform-adopter-001
  secondary_personas: [persona-security-reviewer-002, persona-platform-leader-003]
  affected_profiles: [production-operations, application-support]
situations: [evaluating-adoption, preparing-architecture-review]
content_rules:
  define: [semantic-routing, provider-fallback]
  evidence_required: [architecture-contract, security-control-map, measured-latency]
  objections_to_address: [lock-in, debuggability, central-failure]
  prohibited_claims: [zero-latency, eliminates-provider-lock-in]
  hierarchy: [problem, operating-model, architecture, evidence, failure, adoption]
  primary_cta: run-readiness-assessment
measurement:
  comprehension: [responsibility-model-score]
  decision: [assessment-completion]
  guardrail: [critical-security-misinterpretation-rate]

OKF, linked data, authorship, and agent stewardship

A portable authoring layer with explicit human ownership, source attribution, AI contribution, validation, and semantic derivatives.

The implementation examples strengthen the research architecture: keep durable knowledge separate from agent behavior and memory; validate agent output deterministically; make concepts independently linkable; and record who directed, created, transformed, reviewed, and approved every material artifact.

Adopt OKF as the canonical envelope—not the entire ontology

OKF supplies a portable file contract: Markdown concepts, YAML metadata, ordinary links, progressive index.md files, chronological log.md files, and citations. Our enterprise profile adds research schemas, typed provenance, argument relationships, validation, and JSON-LD compilation while remaining consumable by generic OKF tools.

OKF-SPEC OKF-GOOGLE-BLOG OKF-SKILLS

Operational findings from the examples

Knowledge-as-code: the shared bundle is distinct from agent instructions and session memory.
Deterministic validation: authoring skills are paired with strict conformance checks.
Self-documentation: the toolkit describes its own skills, components, references, and decisions in OKF.
Graph consumption: concepts receive deep links, backlinks, rendered Markdown, and typed metadata views.
Soft-mode upkeep: agents read before work, propose updates after material change, append logs, and validate.
Portable adapters: model-specific skills surround a vendor-neutral canonical knowledge bundle.

OKF-SKILLS OKF-SKILLS-AUTOMATION OKF-SKILLS-SAMPLE OKF-SKILLS-SELF

Our compilation stack

OKF Markdown
Human and agent authoring

JSON Schema
Deterministic profile validation

JSON-LD + PROV
Typed graph and lineage

SPA and outputs
Audience-specific views

Embeddings and search indexes remain disposable derivatives. The files, IDs, metadata, claims, evidence, and provenance remain canonical.

Authorship and contribution model

Agent	What we record	Accountability
Source creator	Author, organization, publisher, date, license, canonical URL, and source origin.	Responsible for the underlying source—not our interpretation.
Jesse Graupmann	Research direction, authorship, editorial judgment, acceptance, and ownership.	Accountable human owner of the synthesis and published outputs.
AI software agent	Provider, product, model, roles, inputs, outputs, time, and review disposition.	Credited contributor and production tool; not the accountable owner.
Reviewer or publisher	Review scope, accepted and rejected changes, approval time, and publication state.	Accountable for the specific review or release decision.

Provenance rule

The work is authored and directed by Jesse Graupmann, with AI working under that direction. AI assistance is visible as leverage rather than obscured or represented as independent ownership. Source authors remain credited for the ideas, evidence, specifications, software, or examples from which the synthesis derives.

Research artifact
  wasAttributedTo → Jesse Graupmann
  wasGeneratedBy  → synthesis activity

Synthesis activity
  wasAssociatedWith → Jesse Graupmann
  wasAssociatedWith → OpenAI ChatGPT / GPT-5.6 Thinking
  used              → cited source entities

AI software agent
  actedOnBehalfOf → Jesse Graupmann

W3C-PROV SCHEMA-CREATIVEWORK JESSE-OKF-VIDEO

Enterprise research concept profile

OKF intentionally permits unknown producer-defined types. These types become our research mechanisms while remaining readable by generic consumers:

Research SourceSource AnnotationResearch ClaimEvidenceWarrantAssumptionConceptInterpretationPersonaSituationAudience Impact ContractArgumentCounterargumentTrade-off AnalysisRecommendationDecisionChallengeProcedureMetricPlaybookOutput ManifestContribution RecordProvenance ActivityGenerated ArtifactSchema ProfileValidation Result

Source schema

Captures original creator, publisher, origin, AI disclosure, evidence class, retrieval date, hash, license, and exact annotations.

Contribution schema

Captures a person, organization, or software agent; its roles; activity; delegation; inputs; outputs; timestamp; and review result.

Concept profile

Adds status, owner, confidence, typed relationships, provenance, and local validation without breaking OKF compatibility.

Governance, red teams, and measurement

The operating controls required to keep research authoritative and useful.

Future-agent red-team sequence

Integrity

Verify sources, versions, hashes, quotations, dates, citation attachment, and retraction or correction status.

Evidence

Test whether each source directly supports the claim and whether scope, causality, and applicability were exaggerated.

Argument

Challenge warrants, assumptions, qualifiers, counterarguments, logical dependencies, and hidden value judgments.

Alternatives

Add status quo, hybrid, phased, build, buy, defer, and abandon options where credible.

Trade-offs

Expose transferred costs, affected stakeholders, uncertainty, reversibility, option value, debt, and lock-in over time.

Audience and ethics

Identify missing personas, affected parties, accessibility constraints, privacy issues, and coercive persuasion.

Freshness and expansion

Find superseded standards, new research, changed enterprise practices, and append a versioned new synthesis.

Agent roles

Evidence auditorMethodologistArgument criticEnterprise strategistFinancial reviewerArchitectSecurity reviewerOperatorPersona advocateAffected-party advocateHistorianFuturistSynthesis editor

Evaluation suite

Test	What better performance means
Citation fidelity	Correctly distinguishes direct support, partial support, context, dispute, and no relationship.
Claim calibration	Narrows or qualifies a claim until it matches the evidence.
Argument completion	Identifies missing warrant, assumption, qualifier, or rebuttal.
Trade-off completeness	Finds omitted costs, beneficiaries, affected parties, horizons, and exit obligations.
Persona rendering	Changes hierarchy and explanation without changing canonical facts.
Temporal update	Compares old and new evidence, preserves history, and proposes a versioned change.
Adversarial synthesis	Produces the strongest evidence-based case against the current recommendation.
Hallucination resistance	Links factual assertions to existing claims and source objects or marks them unsupported.

Outcome measures

Findability · search success · time to answer · comprehension · decision accuracy · task completion · error rate · confidence calibration · accessibility outcome · support deflection · adoption · operational reliability · conversion · internal reuse · citation reuse · freshness · correction rate · retrieval precision.

Measurement boundary

Engagement is not task success. More clicks or time may indicate interest, confusion, or unnecessary effort. Measure the outcome declared in the impact contract.

Atomic research findings

Searchable, evidence-linked conclusions with enterprise implications.

Reader intent and information architecture4 findings

Communication quality is alignment, not prose polish.

The strongest artifact aligns reader intent, evidence strength, decision purpose, information architecture, and delivery medium.

Enterprise implication: Declare purpose, audience, decision, and intended action before drafting.

DIATAXIS HARVARD-ORGANIZING

A universal document creates conflicting contracts.

Tutorials, references, decision memos, specifications, and marketing pages optimize for different questions.

Enterprise implication: Use shared canonical knowledge with separate task-specific views.

DIATAXIS REDHAT-MODULAR

Comprehensive and usable coexist through layering.

Depth belongs in the source layer; progressive disclosure determines what each reader sees first.

Enterprise implication: Preserve full research and publish bounded views.

REDHAT-MODULAR MAYER-MULTIMEDIA

Every substantive unit needs a job.

Paragraphs and sections should advance a claim, explain a concept, provide evidence, support action, or preserve context.

Enterprise implication: Remove or relocate units that serve no declared purpose.

HARVARD-ORGANIZING

Academic and research structures3 findings

Research introductions require a gap and contribution.

Background alone does not establish why new work is needed.

Enterprise implication: Use territory, niche, and contribution moves.

SWALES-CARS

Methods and results should remain distinct from interpretation.

Readers need to inspect how evidence was produced before accepting conclusions.

Enterprise implication: Use IMRaD for experiments, benchmarks, surveys, and formal evaluations.

GMU-IMRAD

Literature reviews synthesize themes rather than narrate a reading list.

Source-by-source summaries transfer analytical work to the reader.

Enterprise implication: Organize agreement, disagreement, methods, limitations, gaps, and implications.

UNC-LIT-REVIEWS

Evidence and argumentation3 findings

Evidence must match the claim.

Standards, experiments, company practices, and professional heuristics support different kinds of conclusions.

Enterprise implication: Label evidence class, scope, confidence, and limitations.

UNC-EVIDENCE

A citation does not automatically support the sentence attached to it.

The source may provide context, a method, partial support, or direct contradiction.

Enterprise implication: Record citation intent and exact source locators.

CITO WEB-ANNOTATION

The warrant is the most frequently hidden part of an enterprise argument.

Evidence does not determine a recommendation without a rule connecting observation to action.

Enterprise implication: Store warrants, backing, qualifiers, and rebuttals as first-class objects.

PURDUE-TOULMIN AIF

Technical documentation and specifications4 findings

Documentation is a product and platform capability.

It has users, journeys, quality attributes, analytics, defects, ownership, release cycles, and deprecation.

Enterprise implication: Assign product ownership, measurement, and lifecycle controls.

SPOTIFY-DOCS-AS-CODE REDHAT-MODULAR

Documentation should follow supported journeys.

Tool-oriented inventories do not necessarily help people complete enterprise work.

Enterprise implication: Organize platform content around Golden Paths, paved roads, and bounded tasks.

SPOTIFY-GOLDEN-PATHS NETFLIX-PAVED-ROADS

A technical specification extends beyond code structure.

Security, privacy, data, performance, failure, observability, rollout, rollback, support, and measures define enterprise readiness.

Enterprise implication: Adopt a complete specification standard with explicit non-goals and alternatives.

MONDAY-TECH-SPEC ATLASSIAN-SDD

ADRs preserve durable decision rationale.

A system design changes; the reasons behind significant choices remain important for future maintainers.

Enterprise implication: Record context, options, decision, consequences, risks, and revisit conditions.

COGNITECT-ADR

Architecture communication3 findings

A diagram is a stakeholder view, not the architecture.

Different concerns require different abstractions, notation, and scope.

Enterprise implication: Declare audience, concern, scope, omissions, legend, and last verification.

IEEE-42010 C4

Architecture diagrams require prose.

Lines and boxes do not reliably communicate ownership, trust, failure, causality, or rationale.

Enterprise implication: Explain purpose, reading order, flows, boundaries, decisions, failures, and operational implications.

IEEE-42010 C4

Context and container views cover many communication needs.

Excessive component or code detail can obscure enterprise boundaries and responsibilities.

Enterprise implication: Start at the highest useful abstraction and add detail only for an explicit question.

Storytelling and presentations3 findings

Story is controlled change, not chronology.

A useful narrative connects an initial state, goal, tension, choice, changed state, and implication.

Enterprise implication: Use narrative to guide attention, then reconnect the case to aggregate evidence.

NARRATIVE-TRANSPORT NARRATIVE-META

A vivid incident does not establish prevalence.

Narratives can be memorable enough to overpower base rates and broader data.

Enterprise implication: Label examples as illustrative and pair them with measured frequency or scope.

NARRATIVE-META

Slides are guided experiences, not repositories.

Dense slides compete with narration and are difficult to reuse independently.

Enterprise implication: Use assertion–evidence slides with notes, appendices, and a companion source document.

ASSERTION-EVIDENCE MAYER-MULTIMEDIA

Marketing, newsletters, and conversion4 findings

Marketing claims must not outrun evidence.

Demonstrated outcomes, customer reports, projections, capabilities, and aspirations have different epistemic status.

Enterprise implication: Classify every external claim before publication.

UNC-EVIDENCE SOURCE-CREDIBILITY

The CTA is a decision interface.

Generic labels can move readers into flows before they understand the consequence.

Enterprise implication: Use action-specific labels appropriate to audience authority and readiness.

NNG-GET-STARTED

A newsletter needs a recurring job.

A list of new content does not create sustained utility.

Enterprise implication: Lead with a signal, explain significance, provide evidence, and use one primary action.

NNG-NEWSLETTER MAILCHIMP-EMAIL

Premium content should add operational leverage.

Gating basic understanding weakens trust and does not establish durable premium value.

Enterprise implication: Monetize depth, proprietary data, tools, benchmarks, support, and customization.

FREEMIUM-VALUE

Personas and audience impact10 findings

A persona is a bounded behavioral decision model.

It is not an average user, demographic stereotype, role, or real participant.

Enterprise implication: State population, decision domain, situation, confidence, and prohibited uses.

MICROSOFT-PERSONAS NNG-PERSONA-TYPES

Persona specificity can reduce representativeness.

Combining many details can create an improbable composite that matches few real people.

Enterprise implication: Include only attributes that materially change a decision or communication need.

CHAPMAN-MILHAM PERSONA-QUANT-TEST

The least fictional persona is usually safer.

Names, photographs, and lifestyle details may improve memory while increasing projection and stereotyping.

Enterprise implication: Prefer archetypes and behavioral labels for enterprise work.

NNG-PERSONAS-ARCHETYPES PERSONA-STEREOTYPE

Persona, situation, and job are separate.

The same senior engineer has different needs while learning, troubleshooting, approving, or presenting.

Enterprise implication: Use a Persona Stack: population → segment → persona → situation → baseline → impact.

DIATAXIS MICROSOFT-PERSONAS

Enterprise artifacts require an audience constellation.

Implementers, approvers, operators, reviewers, affected parties, and future maintainers often have conflicting concerns.

Enterprise implication: Define one primary reading path and layered secondary views.

IEEE-42010 GOVUK-POLICY-PERSONAS

Audience goals should be expressed as impact contracts.

“Inform the audience” is neither observable nor testable.

Enterprise implication: Specify baseline, desired knowledge, decision, behavior, evidence threshold, guardrails, and measure.

WHO-ACTIONABLE IBM-RESEARCH-PLANNING

Persona attributes require provenance.

Observed, reported, measured, inferred, hypothesized, and illustrative data have different validity.

Enterprise implication: Maintain an attribute-level evidence matrix.

MICROSOFT-PERSONAS PERSONA-QUANT-REVIEW

Persona-based recruitment can become circular validation.

Recruiting only people who fit an existing profile cannot reveal missing segments or invalid assumptions.

Enterprise implication: Include contrast cases, holdout samples, disconfirming evidence, and participants who fit none.

NNG-PERSONA-FAIL IBM-RESEARCH-PLANNING

Synthetic personas are not user evidence.

Models can generate plausible language without reproducing population behavior or prevalence.

Enterprise implication: Label synthetic output as hypothesis and validate with real participants and operational data.

LLM-PERSONA-REVIEW LLM-PERSONA-VALIDITY

Participant evidence and persona synthesis must remain separate.

Saying “the persona needs this” can erase the actual sample and variation.

Enterprise implication: Report participant findings first, then link the bounded synthesis.

GOVUK-RESEARCH-PRIVACY MICROSOFT-PERSONAS

Persuasion and persuasive essays10 findings

Responsible persuasion improves informed choice.

The purpose is not agreement at any cost; it is a defensible decision made with material evidence and consequences visible.

Enterprise implication: Evaluate persuasive effectiveness and epistemic integrity separately.

ARISTOTLE-RHETORIC PURDUE-TOULMIN

Logos, ethos, and pathos serve different functions.

Reasoning establishes defensibility, credibility establishes trust, and consequence establishes significance.

Enterprise implication: Let emotional force remain proportional to evidence strength.

ARISTOTLE-RHETORIC

Enterprise communication needs a dual-layer argument.

Decision-makers may scan rapidly but consequential decisions require substantive scrutiny.

Enterprise implication: Lead with decision, stakes, strongest evidence, and trade-off; preserve full methods and analysis.

ELM-REVIEW

Credibility is claim- and context-specific.

Expertise about one product or method does not guarantee independence or applicability elsewhere.

Enterprise implication: Assess competence, integrity, independence, accountability, verifiability, relevance, and currency.

SOURCE-CREDIBILITY

Controlling language can trigger resistance.

Mandates, artificial inevitability, and suppression of alternatives threaten perceived autonomy.

Enterprise implication: Separate mandatory constraints from recommendations, preserve exceptions, and explain reversibility.

REACTANCE

Credible objections should be addressed before approval.

Inoculation and prebunking expose the audience to objections and reasoned responses.

Enterprise implication: Steelman predictable counterarguments and define conditions where they are valid.

INOCULATION

Framing can change the perceived attractiveness of the same option.

Gain, loss, cost, status-quo, and downside frames activate different reference points.

Enterprise implication: Evaluate every material option through multiple equivalent frames.

FRAMING-REVIEW

Explicit uncertainty can support calibrated trust.

Vague uncertainty is less useful than a range, cause, sensitivity, and operational consequence.

Enterprise implication: Separate measurement, forecast, implementation, market, and behavioral uncertainty.

UNCERTAINTY-TRUST

Toulmin is the default structure for practical enterprise recommendations.

Claims are conditional; warrants, qualifiers, and rebuttals determine whether a recommendation actually follows.

Enterprise implication: Expose the complete argument map.

PURDUE-TOULMIN

Rogerian structure is useful for legitimate organizational tension.

Security versus speed and standardization versus autonomy can both contain valid interests.

Enterprise implication: Identify shared goals and create bounded defaults with explicit escape conditions.

PURDUE-ROGERIAN

Business value and temporal trade-offs9 findings

Business value is multidimensional.

Revenue and labor savings omit customer, operational, strategic, technical, workforce, risk, governance, and learning outcomes.

Enterprise implication: Use a complete value model and name the beneficiary.

POLICY-MEMO

Value requires a causal chain.

A capability is not automatically an outcome. Every link from technical change to business consequence needs evidence or a labeled hypothesis.

Enterprise implication: Store outcome → beneficiary → mechanism → measure → baseline → horizon → uncertainty.

PURDUE-TOULMIN

Net value includes transferred and recurring costs.

A proposal may save application-team time while increasing platform, operations, security, procurement, or future-maintainer burden.

Enterprise implication: Name both beneficiary and cost bearer.

POLICY-MEMO SEI-TECH-DEBT

Short- and long-term effects should not share one pros-and-cons list.

Transition disruption, near-term adoption, medium-term operating cost, long-term lock-in, and exit obligations are different analyses.

Enterprise implication: Evaluate immediate, near, medium, long, and exit horizons.

REAL-OPTIONS PATH-DEPENDENCE

Reversibility is both technical and organizational.

A decision can be technically undoable yet economically, contractually, or politically locked in.

Enterprise implication: Estimate exit time, retained obligations, lost data, retraining, and dependency unwinding.

AMAZON-DOORS PATH-DEPENDENCE

Option value can exceed immediate ROI in uncertain environments.

Pilots, interfaces, dual-run capability, export paths, and stop/go milestones preserve future choices.

Enterprise implication: Ask what later decisions become easier or harder.

REAL-OPTIONS

Technical debt is a time-shifted decision.

The metaphor is useful only when the shortcut, benefit, principal, interest, owner, and repayment trigger are explicit.

Enterprise implication: Do not use technical debt as a generic label for disliked code.

SEI-TECH-DEBT

Lock-in is multidimensional.

Data export alone does not address API, identity, telemetry, skill, contract, process, and organizational lock-in.

Enterprise implication: Assess switching cost across every dependency dimension.

PATH-DEPENDENCE

Exploration and exploitation require different measures.

Speculative learning cannot be governed only by the utilization and predictability metrics of mature operations.

Enterprise implication: Classify investments as core scale, experiment, option creation, maintenance, or debt repayment.

AMBIDEXTERITY

LLM retrieval and AEO6 findings

Human-readable structure is machine-useful structure.

Descriptive headings, bounded topics, explicit entities, local context, citations, dates, and versions support both people and retrieval.

Enterprise implication: Design semantic content units before embedding or indexing them.

GOOGLE-AI-FEATURES GOOGLE-WORDS

AEO is authoritative retrievability, not magic markup.

Current search guidance does not require special AI-only schemas in place of people-first content and established indexing foundations.

Enterprise implication: Make claims distinctive, indexable, verifiable, internally linked, and textually available.

GOOGLE-AI-FEATURES GOOGLE-AI-OPT

Chunk at semantic boundaries.

Fixed character windows can split claims from context, evidence, limitations, or procedures.

Enterprise implication: Use concepts, claims, decisions, procedures, failure modes, and examples as retrieval units.

RAG-CHUNKING

Retrieved units require local context.

Pronouns and references such as “this approach” lose meaning when isolated.

Enterprise implication: Use explicit nouns, defined terms, source IDs, scope, and limitations inside each unit.

GOOGLE-WORDS

More context is not necessarily better context.

Large undifferentiated inputs increase noise and can reduce model performance.

Enterprise implication: Retrieve bounded evidence, rerank it, and validate generated output against authority.

AIRBNB-GRAPHQL AIRBNB-VOICE

LLM transformation requires validation.

Models can classify, summarize, restructure, and generate alternate views while introducing unsupported facts or losing boundaries.

Enterprise implication: Validate citations, technical facts, dates, policies, and schemas before publication.

AIRBNB-GRAPHQL LLM-ARCH-DOCS

Research continuity and future agents8 findings

Future agents need research objects, not only final prose.

A final article hides the source graph, reasoning, unresolved questions, and rejected alternatives.

Enterprise implication: Store sources, claims, evidence, warrants, assumptions, recommendations, challenges, and outputs independently.

NANOPUB AIF

A URL is not durable evidence.

Resources change, disappear, or are silently revised.

Enterprise implication: Store canonical and archived URIs, retrieval date, publication date, version, hash, license, and exact locator.

MEMENTO FAIR

Citation intent should be machine-readable.

Future reviewers need to distinguish support, dispute, method reuse, extension, and background.

Enterprise implication: Type each citation relationship.

CITO

Exact source annotations reduce reinterpretation risk.

Document-level references force later agents to rediscover the passage and may invite unsupported paraphrase.

Enterprise implication: Store quotation selectors, prefixes, suffixes, page or section, and associated claim.

WEB-ANNOTATION

Provenance must include LLM activity.

Future researchers need to know which model, prompt, retrieval set, tools, and human review generated a transformation.

Enterprise implication: Represent sources, activities, agents, derivations, approvals, and rejections.

W3C-PROV

Research packages should use open, model-independent formats.

A future model should not require the current vendor, context-window design, or proprietary application.

Enterprise implication: Use Markdown, JSON/YAML with schemas, JSON-LD, CSV/Parquet, text diagrams, IDs, hashes, and tests.

RO-CRATE FAIR

Red-team criticism is a first-class object.

Silently editing a conclusion destroys the history of why confidence changed.

Enterprise implication: Append challenges, resolution status, evidence, and epistemic change logs.

W3C-PROV AIF

Future intelligence needs evaluation tasks.

A more capable model should demonstrate citation fidelity, claim calibration, argument completion, trade-off completeness, persona consistency, and temporal updating.

Enterprise implication: Store test cases with expected criteria, not one frozen answer.

AIRBNB-VOICE FAIR

Enterprise governance and measurement4 findings

Every substantive object needs an owner and lifecycle.

Correct information becomes unsafe when its verification date, scope, or replacement relationship is unknown.

Enterprise implication: Store status, owner, version, review triggers, supersession, and retirement.

FAIR W3C-PROV

Normative language must be controlled.

Must, should, may, and can communicate different authority. Ambiguity creates implementation and audit failures.

Enterprise implication: Define normative terms and separate policy, standard, procedure, and guideline.

WCAG-INFO-REL

Communication success is task success, not engagement alone.

Clicks and time can signal interest, confusion, or friction.

Enterprise implication: Measure comprehension, decision accuracy, task completion, error, retrieval, confidence calibration, and accessibility.

WHO-ACTIONABLE AIRBNB-VOICE

A recommendation requires revisit conditions.

Enterprise decisions are made under bounded evidence and changing environments.

Enterprise implication: Define assumptions, success thresholds, stop conditions, rollback, and review triggers.

REAL-OPTIONS COGNITECT-ADR

OKF, provenance, and agent stewardship9 findings

OKF is an interchange envelope, not a complete epistemic model.

Its minimal Markdown, YAML, file-tree, index, log, link, and citation conventions maximize portability while intentionally leaving domain schemas and relationship semantics to producers.

Enterprise implication: Use OKF for canonical authoring and exchange; add a versioned enterprise research profile, JSON Schema validation, and compiled JSON-LD semantics.

OKF-SPEC OKF-GOOGLE-BLOG

Durable knowledge, agent instructions, and agent memory serve different contracts.

The okf-skills implementation distinguishes shared curated knowledge from standing behavioral instructions and tool-specific implicit memory.

Enterprise implication: Keep research in the OKF bundle, agent behavior in skills or instruction files, and disposable session memory outside the source of truth.

OKF-SKILLS

Agent-generated knowledge still requires deterministic conformance.

The implementation pairs agent authoring with a strict validator rather than relying on an agent to judge its own format compliance.

Enterprise implication: Validate reserved files, YAML parsing, required type fields, local schemas, IDs, internal links, citations, and generated derivatives in CI.

OKF-SKILLS OKF-SPEC

Self-documenting repositories turn documentation into an operational feedback loop.

okf-skills documents its own skills, components, reference specification, and architectural decisions as an OKF bundle that is validated on change.

Enterprise implication: Store the communication research system in the same format it recommends and compile the SPA from that canonical bundle.

OKF-SKILLS-SELF OKF-SKILLS

Concept deep links and backlinks make knowledge navigable beyond the folder tree.

The implementation visualizes typed metadata, outgoing links, citations, and backlinks and gives every concept a shareable fragment deep link.

Enterprise implication: Generate stable concept anchors and backlink indexes in the SPA and graph view; compile typed research relationships into JSON-LD.

OKF-SKILLS OKF-SKILLS-SAMPLE

Soft-mode agent upkeep is safer than silent autonomous rewriting.

The agent template reads relevant knowledge before work, updates affected concepts afterward, appends logs, and runs validation, while deliberately avoiding hidden hooks.

Enterprise implication: Require proposed diffs, contribution records, human review for material changes, and an epistemic log that preserves superseded states.

OKF-SKILLS-AUTOMATION OKF-SKILLS

Source authorship and synthesis contribution are separate provenance dimensions.

The creator of an underlying paper, specification, repository, or video remains the source author even when a human researcher and AI system transform it into a new synthesis.

Enterprise implication: Record source creator and publisher, human research owner, software-agent contribution, derivation activity, review status, and rights separately.

W3C-PROV SCHEMA-CREATIVEWORK JESSE-OKF-VIDEO

AI should be credited as a software-agent contributor without displacing human accountability.

PROV-O can represent a software agent associated with an activity while the resulting entity remains attributed to the accountable human or organization.

Enterprise implication: Use contribution roles such as researched, synthesized, drafted, generated, and validated; identify the human who directed, reviewed, approved, and owns the work.

W3C-PROV SCHEMA-CREATIVEWORK

Public structured data and the internal research graph should be separate projections.

Schema.org JSON-LD should accurately describe visible public content, while the internal graph can express richer provenance, citation intent, claims, challenges, and agent activities.

Enterprise implication: Generate a constrained public JSON-LD block and a downloadable enterprise research graph from the same canonical OKF bundle.

SCHEMA-CREATIVEWORK W3C-PROV OKF-SPEC

Implementation recommendations

Prioritized actions for building a durable enterprise communication and research platform.

P0 establishes integrity and canonical structure. P1 operationalizes the research and communication system. P2 adds argument-aware retrieval, conflicting review agents, formal evaluation, and interoperable packaging.

P0 · Foundation8 recommendations

P0Foundation

Create a canonical research-object repository.

Store sources, claims, evidence, concepts, assumptions, limitations, decisions, challenges, audiences, and outputs separately.

FAIR NANOPUB W3C-PROV

P0Evidence

Require claim-level provenance.

Every material factual claim needs an evidence class, exact source relationship, scope, confidence, limitations, and last verification.

UNC-EVIDENCE CITO WEB-ANNOTATION

P0Argument

Store warrants, qualifiers, and rebuttals.

A recommendation without explicit reasoning cannot be reliably audited or red-teamed.

PURDUE-TOULMIN AIF

P0Audience

Use audience constellations and impact contracts.

Define the primary consumer, decision maker, implementer, operator, reviewer, affected party, and intended measurable outcome.

WHO-ACTIONABLE IEEE-42010

P0Governance

Version knowledge objects instead of overwriting them.

Use status, owner, semantic version, content hash, supersession, review trigger, and epistemic change log.

W3C-PROV FAIR

P0Integrity

Adopt an enterprise persuasion integrity standard.

No recommendation may conceal material alternatives, negative consequences, uncertainty, cost transfer, or evidence limitations.

PURDUE-TOULMIN FRAMING-REVIEW

P0Accessibility

Make semantic structure and accessible meaning non-negotiable.

Use real headings, labels, tables, relationships, keyboard paths, readable contrast, and non-color cues in every output.

WCAG-INFO-REL

P0Preservation

Preserve source states and exact locators.

Store canonical and archive URLs, retrieval date, version, hash, license, and passage-level annotations.

MEMENTO WEB-ANNOTATION

P1 · Core operating model15 recommendations

P1Workflow

Adopt the staged research workflow.

Frame, collect, assess, extract, synthesize, validate, publish, measure, and maintain.

IBM-RESEARCH-PLANNING FAIR

P1Routing

Add a reader-intent router to every project.

Determine whether the primary need is orientation, learning, action, decision, verification, operation, persuasion, or retrieval.

DIATAXIS

P1Technical

Standardize the enterprise technical specification.

Include goals, non-goals, architecture, data, security, failure, observability, capacity, testing, migration, rollout, rollback, ownership, alternatives, and measures.

MONDAY-TECH-SPEC ATLASSIAN-SDD

P1Architecture

Require viewpoint contracts for diagrams.

Each view declares audience, concern, scope, abstraction, omissions, notation, source, owner, and date.

IEEE-42010 C4

P1Persona

Create persona foundation packages, not posters.

Maintain operational card, foundation document, evidence matrix, scenarios, impact contracts, and lifecycle record.

MICROSOFT-PERSONAS

P1Research

Prevent circular persona validation.

Recruit contrast cases, allow people to fit multiple or no personas, and validate against holdout and operational data.

NNG-PERSONA-FAIL PERSONA-QUANT-REVIEW

P1Persuasion

Use Toulmin maps for consequential recommendations.

Expose claim, grounds, warrant, backing, qualifier, rebuttal, assumptions, and alternatives.

PURDUE-TOULMIN

P1Value

Require net-value and causal-chain analysis.

Identify beneficiary, mechanism, measure, baseline, uncertainty, direct cost, operating cost, transition cost, opportunity cost, and cost bearer.

POLICY-MEMO SEI-TECH-DEBT

P1Trade-offs

Use time-expanded trade-off analysis.

Evaluate immediate, near, medium, long-term, and exit effects, including reversibility, option value, debt, and lock-in.

REAL-OPTIONS PATH-DEPENDENCE

P1Presentations

Use assertion–evidence presentation architecture.

Make each slide advance a claim with visual evidence and retain a companion source document.

ASSERTION-EVIDENCE MAYER-MULTIMEDIA

P1Newsletters

Compile newsletters from atomic signal units.

Each item contains a finding, significance, evidence, interpretation, action, and canonical deep link.

NNG-NEWSLETTER MAILCHIMP-EMAIL

P1Marketing

Classify claims before external publication.

Distinguish demonstrated outcome, customer report, internal measurement, model, capability, aspiration, and opinion.

UNC-EVIDENCE SOURCE-CREDIBILITY

P1Retrieval

Build semantic retrieval units.

Chunk by claim, concept, decision, procedure, example, failure mode, and evidence discussion—not fixed character count alone.

RAG-CHUNKING GOOGLE-WORDS

P1Validation

Validate generated outputs against authority.

Check source existence, citation support, dates, technical contracts, normative language, and schema validity.

AIRBNB-GRAPHQL

P1Red team

Create structured challenge and resolution records.

Future agents should append criticism, severity, basis, evidence, status, and proposed resolution.

AIF W3C-PROV

P2 · Advanced knowledge infrastructure8 recommendations

P2Compiler

Build a communication compiler.

Use canonical objects plus persona and output manifests to generate articles, SPAs, decks, specs, newsletters, and retrieval chunks.

REDHAT-MODULAR FAIR

P2Retrieval

Implement argument-aware retrieval.

Retrieve claims with grounds, warrants, scope, limitations, counterarguments, and source annotations.

AIF CITO

P2Agents

Use intentionally conflicting review agents.

Separate evidence auditor, methodologist, argument critic, strategist, architect, security reviewer, operator, persona advocate, and historian.

W3C-PROV

P2Evaluation

Create a future-model evaluation suite.

Test citation fidelity, claim calibration, argument completion, trade-off completeness, persona consistency, temporal updates, and hallucination resistance.

AIRBNB-VOICE FAIR

P2Provenance

Capture full LLM activity provenance.

Record model, version, prompt, retrieval inputs, tool outputs, date, reviewer, accepted changes, rejected changes, and known failures.

W3C-PROV

P2Semantics

Add source annotations and typed citation relationships.

Link exact source passages to claims and state whether a citation supports, disputes, extends, or provides method or background.

WEB-ANNOTATION CITO

P2Analysis

Add scenario and sensitivity modeling.

Show how recommendations change under volume, cost, regulation, adoption, failure, and time-horizon assumptions.

REAL-OPTIONS UNCERTAINTY-TRUST

P2Packaging

Publish a RO-Crate-compatible continuity package.

Bundle README, manifest, source registry, schemas, research objects, provenance, challenges, outputs, evaluations, and changelog.

RO-CRATE FAIR

P0 · OKF and provenance additions3 recommendations

P0Canonical knowledge

Adopt an OKF-compatible bundle as the source of truth.

The SPA, JSON, JSON-LD, search index, and future output formats should be generated derivatives rather than isolated canonical documents.

OKF-SPEC OKF-GOOGLE-BLOG

P0Authorship and provenance

Record source creators, human ownership, and AI contribution separately.

This preserves credit, accountability, derivation, and evidence lineage while showing where AI provided leverage.

W3C-PROV SCHEMA-CREATIVEWORK

P0Validation

Add deterministic OKF, schema, link, and citation checks.

Agents should not be the sole judge of the artifacts they generate. Validation results should be stored and run in CI.

OKF-SKILLS OKF-SPEC

P1 · OKF and provenance additions2 recommendations

P1Agent workflow

Use a soft-mode consume, propose, validate, and review loop.

Agents should consult relevant concepts before work and propose traceable updates afterward without silently rewriting the knowledge base.

OKF-SKILLS-AUTOMATION

P1Navigation

Generate concept deep links, backlinks, and a graph view.

The folder hierarchy supports progressive disclosure; backlinks and typed graph relationships reveal cross-cutting dependencies and evidence lineage.

OKF-SKILLS OKF-SKILLS-SAMPLE

P2 · OKF and provenance additions2 recommendations

P2Semantic exchange

Compile the bundle into an internal JSON-LD and PROV-O graph.

Typed relationships make claims, evidence, derivations, challenges, decisions, and contributions traversable across future systems.

W3C-PROV SCHEMA-CREATIVEWORK

P2Self-documentation

Dogfood the research schema on the research system itself.

The system should contain its own decisions, schemas, build artifacts, validation results, authorship, and update history as first-class concepts.

OKF-SKILLS-SELF

Failure modes and adversarial review

Patterns that make communication look complete while weakening accuracy, usability, or long-term value.

Universal-document failure

Everything in one artifactOnboarding, explanation, API lookup, governance, marketing, and operations compete for hierarchy.
One audience labelA role such as “engineer” hides task, expertise, situation, authority, and consequence of error.
Comprehensiveness as visible densityPreserving source depth is confused with placing all detail in the first view.

Evidence theater

Citation dumpingSources appear without synthesis, exact support, or citation intent.
Authority substitutionA famous company or expert replaces applicability analysis.
Precision theaterExact forecasts imply confidence that input quality does not support.
Confidence launderingA hypothesis or company practice is rewritten as established fact.

Persona fiction

Decorative biographyAge, photograph, hobbies, and family details do not affect the decision.
Circular validationThe persona defines recruitment and the matching sample is used to validate the persona.
Synthetic user evidenceModel-generated responses are reported as participant findings.
Permanent snapshotA persona remains active without evidence-triggered review.

Persuasion without integrity

False urgencyA deadline lacks external constraint or cost-of-delay analysis.
Artificial inevitabilityTrends or executive preference are presented as proof of local fit.
Suppressed objectionsReal alternatives or credible counterarguments are removed.
Emotional asymmetryInaction risks are vivid while action risks remain abstract.
CTA beyond authorityThe reader is asked to approve or implement something they cannot control.

Trade-off compression

Single pros-and-cons listTransition, operating, strategic, and exit consequences are mixed together.
Status-quo omissionThe recommendation is compared only with weak alternatives.
Transferred cost as savingsOne team’s reduced work becomes another team’s uncounted burden.
Technical reversibility claimContract, data, process, skill, and political lock-in are ignored.
Sunk-cost argumentPast investment is treated as a reason to continue regardless of future value.

Architecture ambiguity

Everything diagramMultiple abstraction levels and concerns appear in one unreadable view.
Diagram without narrativeOwnership, trust, failure, and rationale are left to visual inference.
Color-only semanticsMeaning disappears for some readers, print, or alternative rendering.
Unversioned viewThe diagram’s source, owner, and current validity are unknown.

Slide–document hybrid

Paragraph slidesThe audience reads while the speaker narrates competing information.
Topic-title deckSlides label categories but make no assertions.
Deck as source of truthMethods, caveats, and citations disappear when the speaker is absent.

Marketing overreach

Capability becomes outcomeThe causal chain to business value is not shown.
Seamless and zero-riskUnmeasurable absolutes replace conditions and constraints.
Understanding is gatedThe reader must register or pay before understanding the offer.
Generic CTALearn more or get started obscures what happens next.

LLM knowledge debt

Final prose onlyFuture models cannot audit sources, warrants, challenges, or missing evidence.
Embeddings as canonical dataAn index replaces human-readable and structured source material.
Unlimited contextLarge noisy input replaces targeted retrieval and validation.
Silent model transformationPrompt, model, retrieval set, and reviewer are not recorded.

Silent history loss

Overwrite instead of supersedePrior claims and their context disappear.
Rejected challenge deletionFuture reviewers cannot see what was considered and why it was rejected.
Text-only changelogThe wording changed, but the knowledge change is not explained.
Unowned living documentNo person or team is responsible for verification and retirement.

Reference registry

Evidence grouped by concept with guidance for how each source should be used.

documentation

DIATAXISDiátaxis documentation frameworkReader intent and the separation of tutorials, how-to guides, reference, and explanation.Framework

Use for: Reader intent and the separation of tutorials, how-to guides, reference, and explanation.
Reference as: Use distinct documentation forms for distinct user needs.
Best application: Documentation architecture and content audits.
Evidence class: Framework

Open source ↗

REDHAT-MODULARRed Hat modular documentationConcept, procedure, and reference modules that answer bounded user questions.Practice

Use for: Concept, procedure, and reference modules that answer bounded user questions.
Reference as: Modular content can be composed into different experiences.
Best application: Docs-as-products and reusable content systems.
Evidence class: Practice

Open source ↗

GOOGLE-HEADINGSGoogle developer documentation: HeadingsDescriptive, unique, hierarchical headings.Practice

Use for: Descriptive, unique, hierarchical headings.
Reference as: Headings should expose purpose and structure.
Best application: Technical documentation and retrieval-friendly pages.
Evidence class: Practice

Open source ↗

GOOGLE-WORDSGoogle technical writing: WordsConsistent terminology, explicit nouns, and restrained acronyms.Practice

Use for: Consistent terminology, explicit nouns, and restrained acronyms.
Reference as: Local clarity improves human and machine interpretation.
Best application: Documentation, RAG chunks, and specs.
Evidence class: Practice

Open source ↗

accessibility

WCAG-INFO-RELWCAG 2.2: Info and RelationshipsSemantic representation of structure and relationships.Standard

Use for: Semantic representation of structure and relationships.
Reference as: Presentation alone must not carry essential structure.
Best application: HTML, documents, tables, diagrams, and machine retrieval.
Evidence class: Standard

Open source ↗

academic

HARVARD-ORGANIZINGHarvard: Organizing Your EssayThesis decomposition, paragraph purpose, evidence, and analysis.University

Use for: Thesis decomposition, paragraph purpose, evidence, and analysis.
Reference as: A strong argument advances through explicit claims and subclaims.
Best application: Research articles, memos, and persuasive writing.
Evidence class: University

Open source ↗

UNC-EVIDENCEUNC: EvidenceDiscipline-appropriate forms of evidence and their use.University

Use for: Discipline-appropriate forms of evidence and their use.
Reference as: Evidence must match the kind of claim being made.
Best application: Evidence policies and claim assessment.
Evidence class: University

Open source ↗

UNC-LIT-REVIEWSUNC: Literature ReviewsLiterature reviews as synthesis rather than source-by-source summaries.University

Use for: Literature reviews as synthesis rather than source-by-source summaries.
Reference as: Organize existing work by themes, debates, methods, and gaps.
Best application: Deep-research reports and source maps.
Evidence class: University

Open source ↗

UNC-ARGUMENTUNC: ArgumentClaims, evidence, reasoning, and counterargument.University

Use for: Claims, evidence, reasoning, and counterargument.
Reference as: Facts become arguments only when connected by reasoning.
Best application: Recommendations and technical position papers.
Evidence class: University

Open source ↗

GMU-IMRADGeorge Mason: IMRaD ReportsIntroduction, methods, results, and discussion.University

Use for: Introduction, methods, results, and discussion.
Reference as: Separate observation from method and interpretation.
Best application: Experiments, benchmarks, surveys, and evaluations.
Evidence class: University

Open source ↗

SWALES-CARSSwales CARS modelEstablish territory, establish niche, occupy niche.Primary

Use for: Establish territory, establish niche, occupy niche.
Reference as: Introductions should show importance, unresolved need, and contribution.
Best application: Research introductions, proposals, and thought leadership.
Evidence class: Primary

Open source ↗

technical

MONDAY-TECH-SPECMonday.com: Technical specificationFunctional versus technical specifications, rollout, security, support, and metrics.Practice

Use for: Functional versus technical specifications, rollout, security, support, and metrics.
Reference as: Specifications must connect implementation to business and operations.
Best application: Enterprise technical-spec templates.
Evidence class: Practice

Open source ↗

ATLASSIAN-SDDAtlassian: Software design documentArchitecture, interfaces, assumptions, dependencies, constraints, and trade-offs.Practice

Use for: Architecture, interfaces, assumptions, dependencies, constraints, and trade-offs.
Reference as: Design documents are shared alignment and decision artifacts.
Best application: Design reviews and implementation planning.
Evidence class: Practice

Open source ↗

enterprise

AMAZON-NARRATIVESAWS: Product management at AmazonNarrative mechanisms including PR/FAQ, reviews, readiness, and correction of error.Practice

Use for: Narrative mechanisms including PR/FAQ, reviews, readiness, and correction of error.
Reference as: Different narrative forms serve different organizational decisions.
Best application: Decision memos and operational reviews.
Evidence class: Practice

Open source ↗

POLICY-MEMOPolicy memo guidanceDecision-maker focus, evidence, alternatives, feasibility, and recommendation.University

Use for: Decision-maker focus, evidence, alternatives, feasibility, and recommendation.
Reference as: Enterprise recommendations should be decision instruments, not generic essays.
Best application: Executive decision papers.
Evidence class: University

Open source ↗

architecture

COGNITECT-ADRDocumenting Architecture DecisionsContext, decision, status, and consequences for significant architecture choices.Framework

Use for: Context, decision, status, and consequences for significant architecture choices.
Reference as: Preserve why a decision was made, including negative consequences.
Best application: Architecture decision records.
Evidence class: Framework

Open source ↗

IEEE-42010ISO/IEC/IEEE 42010 architecture descriptionsStakeholders, concerns, viewpoints, and views.Standard

Use for: Stakeholders, concerns, viewpoints, and views.
Reference as: One diagram cannot represent every architectural concern.
Best application: Architecture documentation systems.
Evidence class: Standard

Open source ↗

C4C4 modelHierarchical system, container, component, and code views.Framework

Use for: Hierarchical system, container, component, and code views.
Reference as: Use progressive architectural zoom for different audiences.
Best application: System and platform explainers.
Evidence class: Framework

Open source ↗

platform

SPOTIFY-GOLDEN-PATHSSpotify Golden PathsSupported journeys, tutorials, tooling, and platform enablement.Practice

Use for: Supported journeys, tutorials, tooling, and platform enablement.
Reference as: Documentation works best when attached to supported product paths.
Best application: Platform-engineering documentation.
Evidence class: Practice

Open source ↗

SPOTIFY-DOCS-AS-CODESpotify docs as code and BackstageDocumentation close to code and changed through engineering workflows.Practice

Use for: Documentation close to code and changed through engineering workflows.
Reference as: Documentation ownership should follow software ownership.
Best application: Technical-doc lifecycle design.
Evidence class: Practice

Open source ↗

NETFLIX-PAVED-ROADSNetflix paved roadsSupported practices and tools made easier than unsupported alternatives.Practice

Use for: Supported practices and tools made easier than unsupported alternatives.
Reference as: Standardization requires productized enablement, not prose alone.
Best application: Platform adoption strategy.
Evidence class: Practice

Open source ↗

AIRBNB-VIADUCTAirbnb Viaduct documentationRole- and task-separated documentation, API reference, RFCs, and stability annotations.Practice

Use for: Role- and task-separated documentation, API reference, RFCs, and stability annotations.
Reference as: Organize platform docs by audience journey and lifecycle.
Best application: Platform documentation information architecture.
Evidence class: Practice

Open source ↗

llm

AIRBNB-GRAPHQLAirbnb: GraphQL data mocking with LLMsBounded relevant context, documentation, schema validation, and corrective retries.Practice

Use for: Bounded relevant context, documentation, schema validation, and corrective retries.
Reference as: Provide models only relevant context and validate output against authority.
Best application: RAG and agentic documentation systems.
Evidence class: Practice

Open source ↗

AIRBNB-VOICEAirbnb voice support retrievalSemantic retrieval, reranking, and retrieval-quality measurement.Practice

Use for: Semantic retrieval, reranking, and retrieval-quality measurement.
Reference as: Retrieval systems require evaluation, not only generation quality.
Best application: Enterprise knowledge retrieval.
Evidence class: Practice

Open source ↗

story

NARRATIVE-TRANSPORTNarrative transportation researchAttention, imagery, emotion, and immersion in narratives.Review

Use for: Attention, imagery, emotion, and immersion in narratives.
Reference as: Story guides attention but can influence beyond evidence strength.
Best application: Case studies and technical storytelling.
Evidence class: Review

Open source ↗

DATA-STORY-REVIEWData storytelling systematic reviewNarrative, visualization, cognition, and interaction in data storytelling.Review

Use for: Narrative, visualization, cognition, and interaction in data storytelling.
Reference as: Data storytelling is broader than a single narrative formula.
Best application: Research visualization and analytical communication.
Evidence class: Review

Open source ↗

presentation

ASSERTION-EVIDENCEAssertion–evidence presentationsComplete-sentence assertions supported by visual evidence.University

Use for: Complete-sentence assertions supported by visual evidence.
Reference as: Slides should advance claims rather than display topic headings and bullet walls.
Best application: Technical and executive presentations.
Evidence class: University

Open source ↗

MAYER-MULTIMEDIAMultimedia learning principlesCoherence, signaling, redundancy, and spatial/temporal contiguity.Review

Use for: Coherence, signaling, redundancy, and spatial/temporal contiguity.
Reference as: Reduce extraneous processing and align related verbal and visual information.
Best application: Slides, diagrams, and interactive explainers.
Evidence class: Review

Open source ↗

marketing

AIDAAIDA model reviewAttention, interest, desire, and action as a historical persuasion heuristic.Review

Use for: Attention, interest, desire, and action as a historical persuasion heuristic.
Reference as: AIDA is a drafting aid, not a universal linear decision model.
Best application: Marketing hierarchy with qualification.
Evidence class: Review

Open source ↗

NNG-GET-STARTEDNN/g: Get Started linksGeneric CTAs can pull users into flows before they understand the offer.Practice

Use for: Generic CTAs can pull users into flows before they understand the offer.
Reference as: CTA labels should describe the consequence of action.
Best application: Landing pages and calls to action.
Evidence class: Practice

Open source ↗

FREEMIUM-VALUEFreemium satisfaction and conversion researchPerceived value and satisfaction in premium conversion.Primary

Use for: Perceived value and satisfaction in premium conversion.
Reference as: Premium value should add utility rather than gate basic understanding.
Best application: Research-product packaging, with context limits.
Evidence class: Primary

Open source ↗

MAILCHIMP-EMAILMailchimp email marketing designClear goal, concise content, CTA, responsive design, and testing.Practice

Use for: Clear goal, concise content, CTA, responsive design, and testing.
Reference as: Email effectiveness must be tested with the actual audience.
Best application: Newsletter and campaign design.
Evidence class: Practice

Open source ↗

persona

MICROSOFT-PERSONASMicrosoft personas in practice and theoryFoundation documents, traceability, scenarios, progressive disclosure, and revision.Primary

Use for: Foundation documents, traceability, scenarios, progressive disclosure, and revision.
Reference as: Personas should be maintained research infrastructure.
Best application: Enterprise persona programs.
Evidence class: Primary

Open source ↗

PERSONA-QUANT-REVIEWReview of quantitative persona creationRigor, scalability, objectivity, representation, and mixed methods.Review

Use for: Rigor, scalability, objectivity, representation, and mixed methods.
Reference as: Use qualitative discovery with quantitative validation where possible.
Best application: Persona methodology.
Evidence class: Review

Open source ↗

CHAPMAN-MILHAMPersonas and verification critiqueVerification, falsifiability, and population representation concerns.Primary

Use for: Verification, falsifiability, and population representation concerns.
Reference as: A concrete profile may not represent a measurable segment.
Best application: Persona red-teaming.
Evidence class: Primary

Open source ↗

PERSONA-QUANT-TESTQuantitative test of persona specificityMatch rates decline as many attributes are combined.Primary

Use for: Match rates decline as many attributes are combined.
Reference as: Only include decision-relevant attributes.
Best application: Persona schema design.
Evidence class: Primary

Open source ↗

PERSONA-STEREOTYPEPersona stereotyping critiqueRisks of simplification and stereotyping in persona representations.Primary

Use for: Risks of simplification and stereotyping in persona representations.
Reference as: Humanizing details can increase projection and exclusion.
Best application: Ethical persona review.
Evidence class: Primary

Open source ↗

NNG-REVISE-PERSONASNN/g: Revising personasPersonas drift as products, behavior, and environments change.Practice

Use for: Personas drift as products, behavior, and environments change.
Reference as: Use evidence-triggered review and versioning.
Best application: Persona lifecycle.
Evidence class: Practice

Open source ↗

NNG-PERSONA-TYPESNN/g: Persona typesProto-personas, qualitative personas, and statistically supported personas.Practice

Use for: Proto-personas, qualitative personas, and statistically supported personas.
Reference as: Label the evidence maturity of every persona.
Best application: Persona governance.
Evidence class: Practice

Open source ↗

NNG-PERSONAS-ARCHETYPESNN/g: Personas versus archetypesBiographical representation versus abstract behavioral patterns.Practice

Use for: Biographical representation versus abstract behavioral patterns.
Reference as: Use the least fictional form that supports the decision.
Best application: Enterprise audience modeling.
Evidence class: Practice

Open source ↗

WHO-ACTIONABLEWHO actionable communicationAudience knowledge, attitudes, behavior, barriers, and action.Practice

Use for: Audience knowledge, attitudes, behavior, barriers, and action.
Reference as: Communication objectives should specify observable audience outcomes.
Best application: Persona impact contracts.
Evidence class: Practice

Open source ↗

NNG-PERSONA-FAILNN/g: Why personas failScope, organizational embedding, and connection to decisions.Practice

Use for: Scope, organizational embedding, and connection to decisions.
Reference as: A persona poster without decision integration is decoration.
Best application: Persona program review.
Evidence class: Practice

Open source ↗

GOVUK-POLICY-PERSONASGOV.UK policy persona guidanceSubstantial research, participation, observed facts, and continuing reevaluation.Practice

Use for: Substantial research, participation, observed facts, and continuing reevaluation.
Reference as: Represent affected people, not only interface users.
Best application: Policy and service personas.
Evidence class: Practice

Open source ↗

ONS-PERSONASONS content personasAudience grouping by expertise and task.Practice

Use for: Audience grouping by expertise and task.
Reference as: Communication personas should emphasize knowledge and use context.
Best application: Content design.
Evidence class: Practice

Open source ↗

ATLASSIAN-BUYER-PERSONASAtlassian buyer personasBuying role, channels, trusted sources, goals, and barriers.Practice

Use for: Buying role, channels, trusted sources, goals, and barriers.
Reference as: Buyer and product-use personas support different decisions.
Best application: Enterprise marketing.
Evidence class: Practice

Open source ↗

MICROSOFT-PERSONA-POWERMicrosoft: The power of personasPersona use boundaries and qualitative judgment.Practice

Use for: Persona use boundaries and qualitative judgment.
Reference as: Do not use persona opinion when hard quantitative criteria are required.
Best application: Persona governance.
Evidence class: Practice

Open source ↗

LLM-PERSONA-REVIEWSystematic review of LLM-generated personasGrowth, evaluation gaps, and human oversight for synthetic personas.Review

Use for: Growth, evaluation gaps, and human oversight for synthetic personas.
Reference as: Synthetic personas are hypotheses, not empirical participants.
Best application: LLM persona policies.
Evidence class: Review

Open source ↗

LLM-PERSONA-VALIDITYPersona prompting and subgroup validityEmerging evidence that persona conditioning may not reproduce population behavior.Emerging

Use for: Emerging evidence that persona conditioning may not reproduce population behavior.
Reference as: Do not treat simulated responses as user research.
Best application: Synthetic-user red teams.
Evidence class: Draft

Open source ↗

research

IBM-RESEARCH-PLANNINGIBM research planningObjectives, business goals, participant groups, recruitment, methods, and repositories.Practice

Use for: Objectives, business goals, participant groups, recruitment, methods, and repositories.
Reference as: Persona research belongs inside a formal research plan.
Best application: Research operations.
Evidence class: Practice

Open source ↗

GOVUK-RESEARCH-PRIVACYGOV.UK participant privacyConsent, data minimization, controlled access, and deletion.Standard

Use for: Consent, data minimization, controlled access, and deletion.
Reference as: Separate restricted participant data from aggregated persona outputs.
Best application: Research privacy governance.
Evidence class: Standard

Open source ↗

persuasion

ARISTOTLE-RHETORICStanford Encyclopedia: Aristotle rhetoricLogos, ethos, and pathos in rhetorical persuasion.Review

Use for: Logos, ethos, and pathos in rhetorical persuasion.
Reference as: Reasoning, credibility, and significance must reinforce one another.
Best application: Enterprise persuasive structure.
Evidence class: Review

Open source ↗

ELM-REVIEWElaboration Likelihood Model reviewMotivation, ability, elaboration, arguments, and cues.Review

Use for: Motivation, ability, elaboration, arguments, and cues.
Reference as: Layer enterprise decisions for both rapid orientation and substantive scrutiny.
Best application: Executive and technical dual-layer content.
Evidence class: Review

Open source ↗

SOURCE-CREDIBILITYSource credibility research reviewCredibility, ambiguity, expertise, and audience evaluation.Review

Use for: Credibility, ambiguity, expertise, and audience evaluation.
Reference as: Credibility applies per claim and context, not universally to a brand.
Best application: Evidence assessment.
Evidence class: Review

Open source ↗

REACTANCEPsychological reactance reviewResistance when freedom is perceived as threatened.Review

Use for: Resistance when freedom is perceived as threatened.
Reference as: Separate mandates from recommendations and preserve meaningful choice.
Best application: Adoption communication.
Evidence class: Review

Open source ↗

INOCULATIONMeta-analysis of inoculation theoryExposure to counterarguments and refutations.Review

Use for: Exposure to counterarguments and refutations.
Reference as: Address credible objections before they become external attacks.
Best application: Decision papers and prebunking.
Evidence class: Review

Open source ↗

NARRATIVE-METANarrative persuasion meta-analysisVariation in narrative effects by audience, topic, medium, and familiarity.Review

Use for: Variation in narrative effects by audience, topic, medium, and familiarity.
Reference as: Use stories to make consequences concrete, then return to aggregate evidence.
Best application: Case studies and presentations.
Evidence class: Review

Open source ↗

FRAMING-REVIEWFraming effects reviewJudgments change across gain, loss, and reference-point frames.Review

Use for: Judgments change across gain, loss, and reference-point frames.
Reference as: Test recommendations in benefit, loss, cost, status-quo, and downside frames.
Best application: Decision quality.
Evidence class: Review

Open source ↗

UNCERTAINTY-TRUSTUncertainty communication and trustTrust effects of explicit and quantified uncertainty.Primary

Use for: Trust effects of explicit and quantified uncertainty.
Reference as: Explain uncertainty type, range, driver, and consequence.
Best application: Forecasts and recommendations.
Evidence class: Primary

Open source ↗

PURDUE-TOULMINPurdue OWL: Toulmin argumentClaim, grounds, warrant, backing, qualifier, and rebuttal.University

Use for: Claim, grounds, warrant, backing, qualifier, and rebuttal.
Reference as: Make practical reasoning inspectable and challengeable.
Best application: Enterprise recommendations.
Evidence class: University

Open source ↗

PURDUE-CLASSICALPurdue OWL: Classical argumentContext, position, proof, refutation, and conclusion.University

Use for: Context, position, proof, refutation, and conclusion.
Reference as: Use a broad persuasive arc for strategic advocacy.
Best application: Articles and presentations.
Evidence class: University

Open source ↗

PURDUE-ROGERIANPurdue OWL: Rogerian argumentFair presentation of opposing positions and common ground.University

Use for: Fair presentation of opposing positions and common ground.
Reference as: Use when legitimate enterprise interests are in tension.
Best application: Centralization, autonomy, and governance disputes.
Evidence class: University

Open source ↗

tradeoff

AMAZON-DOORSAmazon one-way and two-way doorsDecision-process intensity based on reversibility.Practice

Use for: Decision-process intensity based on reversibility.
Reference as: Hard-to-reverse decisions deserve stronger evidence and review.
Best application: Decision governance, as a company heuristic.
Evidence class: Practice

Open source ↗

REAL-OPTIONSReal options reasoningExpansion, deferral, switching, and abandonment under uncertainty.Primary

Use for: Expansion, deferral, switching, and abandonment under uncertainty.
Reference as: Evaluate which future choices an investment creates or closes.
Best application: AI and platform strategy.
Evidence class: Primary

Open source ↗

SEI-TECH-DEBTSEI: Field study of technical debtShort-term delivery versus future evolution and architectural debt.Primary

Use for: Short-term delivery versus future evolution and architectural debt.
Reference as: Document principal, interest, benefit, trigger, and owner.
Best application: Architecture trade-offs.
Evidence class: Primary

Open source ↗

PATH-DEPENDENCETechnology path dependencePositive feedback, accumulated investment, and lock-in.Primary

Use for: Positive feedback, accumulated investment, and lock-in.
Reference as: Portability must be evaluated across data, API, skills, contracts, and operations.
Best application: Long-term option analysis.
Evidence class: Primary

Open source ↗

AMBIDEXTERITYOrganizational ambidexterity reviewExploration versus exploitation and organizational structure.Review

Use for: Exploration versus exploitation and organizational structure.
Reference as: Separate mature operations from bounded experimentation and strategic learning.
Best application: Portfolio recommendations.
Evidence class: Review

Open source ↗

stewardship

FAIRFAIR principlesFindability, accessibility, interoperability, and reuse.Standard

Use for: Findability, accessibility, interoperability, and reuse.
Reference as: Research needs identifiers, metadata, provenance, relationships, and licenses.
Best application: Future-proof research packages.
Evidence class: Standard

Open source ↗

W3C-PROVW3C PROV-OEntities, activities, agents, derivation, attribution, and revision.Standard

Use for: Entities, activities, agents, derivation, attribution, and revision.
Reference as: Capture how research objects and outputs were generated.
Best application: Human and LLM provenance.
Evidence class: Standard

Open source ↗

MEMENTORFC 7089 MementoTime-based access to archived states of web resources.Standard

Use for: Time-based access to archived states of web resources.
Reference as: A URL alone is insufficient for durable evidence.
Best application: Source preservation.
Evidence class: Standard

Open source ↗

DATASHEETSDatasheets for DatasetsMotivation, composition, collection, use, and limitations documentation.Primary

Use for: Motivation, composition, collection, use, and limitations documentation.
Reference as: Apply structured documentation cards to research sources and syntheses.
Best application: Research-object records.
Evidence class: Primary

Open source ↗

NANOPUBNanopublication guidelinesAtomic assertions packaged with provenance and publication information.Framework

Use for: Atomic assertions packaged with provenance and publication information.
Reference as: Claims should be independently identifiable and citable.
Best application: Claim ledgers.
Evidence class: Framework

Open source ↗

AIFArgument Interchange FormatShared representation for structured arguments.Primary

Use for: Shared representation for structured arguments.
Reference as: Store the reasoning graph, not only the final recommendation.
Best application: Argument-aware research systems.
Evidence class: Primary

Open source ↗

CITOCitation Typing OntologyMachine-readable citation intent such as supports, disputes, or extends.Primary

Use for: Machine-readable citation intent such as supports, disputes, or extends.
Reference as: Record why each source is cited.
Best application: Citation graph design.
Evidence class: Primary

Open source ↗

WEB-ANNOTATIONW3C Web Annotation Data ModelAnnotations linked to exact text or resource segments.Standard

Use for: Annotations linked to exact text or resource segments.
Reference as: Preserve source locators, quotations, and evidence relationships.
Best application: Evidence extraction.
Evidence class: Standard

Open source ↗

RO-CRATERO-Crate specification and guidanceLightweight JSON-LD research-object packaging.Standard

Use for: Lightweight JSON-LD research-object packaging.
Reference as: Use an open outer container for files, metadata, people, and software.
Best application: Future-agent handoff packages.
Evidence class: Standard

Open source ↗

ORCIDORCID persistent researcher identifiersPersistent identities for research contributors.Standard

Use for: Persistent identities for research contributors.
Reference as: Keep agent and human attribution stable over time.
Best application: Research provenance.
Evidence class: Standard

Open source ↗

SWHIDSoftware Heritage persistent identifiersContent-based persistent identifiers for software artifacts.Standard

Use for: Content-based persistent identifiers for software artifacts.
Reference as: Use hashes and stable logical IDs for exact artifact states.
Best application: Code and research integrity.
Evidence class: Standard

Open source ↗

retrieval

GOOGLE-AI-FEATURESGoogle AI features and website contentIndexable, people-first, internally linked, textually available content.Practice

Use for: Indexable, people-first, internally linked, textually available content.
Reference as: No special AI markup replaces clear authoritative content.
Best application: AEO and discoverability.
Evidence class: Practice

Open source ↗

GOOGLE-AI-OPTGoogle guidance on AI and search contentExisting quality and SEO foundations for AI-mediated discovery.Practice

Use for: Existing quality and SEO foundations for AI-mediated discovery.
Reference as: Do not rewrite content into artificial machine-targeted keyword patterns.
Best application: AEO governance.
Evidence class: Practice

Open source ↗

RAG-CHUNKINGStructure-aware RAG chunking researchHierarchical and structure-aware segmentation in retrieval systems.Emerging

Use for: Hierarchical and structure-aware segmentation in retrieval systems.
Reference as: Chunk at semantic boundaries and validate against the actual corpus.
Best application: RAG experimentation.
Evidence class: Draft

Open source ↗

LLM-ARCH-DOCSLLM-generated architecture documentationEmerging value and limitations of automated architecture documentation.Emerging

Use for: Emerging value and limitations of automated architecture documentation.
Reference as: Use LLMs for transformation with technical validation and governance.
Best application: Documentation automation.
Evidence class: Draft

Open source ↗

OKF, authorship, and provenance8 sources

OKF-SPECOpen Knowledge Format (OKF) Version 0.1 — DraftCanonical rules for bundles, concepts, frontmatter, cross-links, index files, logs, citations, conformance, and versioning.Standard

Use for: Canonical rules for bundles, concepts, frontmatter, cross-links, index files, logs, citations, conformance, and versioning.
Reference as: OKF defines a deliberately minimal interoperability envelope rather than a centrally registered domain ontology.
Best application: Canonical bundle format and conformance baseline.
Creator / author: Google Cloud Data Cloud team
Publisher: Google Cloud
Origin: human and organizational specification

Open source ↗

OKF-GOOGLE-BLOGIntroducing the Open Knowledge FormatOfficial motivation and design principles: minimal opinion, producer/consumer independence, and format rather than platform.Practice

Use for: Official motivation and design principles: minimal opinion, producer/consumer independence, and format rather than platform.
Reference as: OKF formalizes a portable LLM-wiki pattern for human and agent consumption without requiring a proprietary runtime or SDK.
Best application: Enterprise adoption rationale.
Creator / author: Sam McVeety, Amir Hormati
Publisher: Google Cloud
Origin: human-authored first-party practice

Open source ↗

OKF-SKILLSokf-skills — OKF toolkit for Claude Code and agent skillsAgent production, maintenance, strict validation, visualization, deep links, portable skill distribution, and knowledge-as-code layering.Practice

Use for: Agent production, maintenance, strict validation, visualization, deep links, portable skill distribution, and knowledge-as-code layering.
Reference as: A community implementation demonstrates deterministic validation and a self-contained graph consumer around the minimal OKF format.
Best application: Agent workflow and tooling patterns.
Creator / author: Marco Boffo (scaccogatto)
Publisher: scaccogatto
Origin: human-authored open-source software

Open source ↗

OKF-SKILLS-AUTOMATIONCLAUDE-okf soft-mode upkeep templateRead index before relevant work, update affected concepts and logs after change, then validate strictly.Practice

Use for: Read index before relevant work, update affected concepts and logs after change, then validate strictly.
Reference as: Agent upkeep can be explicitly instructed and validated without hidden hooks or treating memory as the source of truth.
Best application: Future-agent handoff and update workflow.
Creator / author: Marco Boffo (scaccogatto)
Publisher: scaccogatto
Origin: human-authored agent instruction

Open source ↗

OKF-SKILLS-SAMPLEStorefront sample OKF bundleConcrete bundle containing services, datasets, decisions, runbooks, metrics, index navigation, and a graph visualization.Practice

Use for: Concrete bundle containing services, datasets, decisions, runbooks, metrics, index navigation, and a graph visualization.
Reference as: Different concept types can coexist in one linked bundle and be rendered into a shareable self-contained explorer.
Best application: Bundle organization and concept-type examples.
Creator / author: Marco Boffo (scaccogatto)
Publisher: scaccogatto
Origin: human-authored open-source example

Open source ↗

OKF-SKILLS-SELFokf-skills documented in its own OKF bundleSelf-documenting repository with skills, components, reference specifications, and architecture decisions.Practice

Use for: Self-documenting repository with skills, components, reference specifications, and architecture decisions.
Reference as: Dogfooding creates a feedback loop in which the knowledge system documents and validates its own design.
Best application: Self-documentation and architecture lineage.
Creator / author: Marco Boffo (scaccogatto)
Publisher: scaccogatto
Origin: human-authored open-source example

Open source ↗

JESSE-OKF-VIDEOOKF usage example and AI-assisted research workflow (video)First-party example supplied by the research owner demonstrating how OKF can be applied and communicated.Practice

Use for: First-party example supplied by the research owner demonstrating how OKF can be applied and communicated.
Reference as: Track the originating human creator and disclose AI assistance as separate provenance fields.
Best application: Authorship, ownership, and AI-assistance example.
Creator / author: Jesse Graupmann
Publisher: YouTube
Origin: human-directed, AI-assisted
Verification: Source and creator attribution supplied by the research owner; transcript was not independently captured in this update.

Open source ↗

SCHEMA-CREATIVEWORKSchema.org CreativeWork and SoftwareApplicationPublic structured-data properties for creator, contributor, credit, copyright, dates, versions, citations, and derivation.Standard

Use for: Public structured-data properties for creator, contributor, credit, copyright, dates, versions, citations, and derivation.
Reference as: Use Schema.org to describe the public artifact; use PROV-O for richer software-agent activities and derivation.
Best application: Public JSON-LD and attribution.
Creator / author: Schema.org Community Group
Publisher: Schema.org
Origin: community vocabulary

Open source ↗

LLM and machine-readable data

The visible interface and embedded data share the same research model.

Embedded dataset

The page contains structured metadata, principles, domains, genres, structures, findings, playbooks, recommendations, anti-patterns, value dimensions, OKF profile, authorship, contribution records, object types, schemas, reference records, JSON-LD provenance, and future-agent handoff rules.

document.getElementById('research-data').textContent

Agent constraints

Do not delete prior claims.
Do not convert inference into fact.
Do not upgrade confidence without evidence.
Do not treat synthetic personas as participant evidence.
Do not silently repair or replace citations.
Record contradictions, proposed changes, confidence, provenance, human ownership, and software-agent contributions.
Validate OKF and local schemas deterministically before publication.

Research source and authorshipOriginal creator, publisher, source origin, AI disclosure, rights, and integrity metadata.

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://example.org/schemas/research-source.schema.json",
  "title": "Research Source Record",
  "type": "object",
  "required": ["id", "title", "canonical_url", "source_type", "attribution_status"],
  "properties": {
    "id": {"type": "string"},
    "title": {"type": "string"},
    "canonical_url": {"type": "string", "format": "uri"},
    "source_type": {"type": "string"},
    "creators": {"type": "array", "items": {"$ref": "#/$defs/agent"}},
    "publisher": {"$ref": "#/$defs/agent"},
    "authorship_origin": {"enum": ["human", "organization", "human-directed-ai-assisted", "ai-generated", "mixed", "unknown"]},
    "attribution_status": {"enum": ["verified", "publisher-only", "pending", "unknown"]},
    "ai_disclosure": {"type": ["object", "null"]},
    "published_at": {"type": ["string", "null"]},
    "retrieved_at": {"type": "string"},
    "content_hash": {"type": ["string", "null"]},
    "license": {"type": ["string", "null"]}
  },
  "$defs": {
    "agent": {
      "type": "object",
      "required": ["name", "type"],
      "properties": {
        "id": {"type": ["string", "null"]},
        "name": {"type": "string"},
        "type": {"enum": ["Person", "Organization", "SoftwareAgent"]},
        "role": {"type": ["string", "null"]}
      }
    }
  }
}

Human and AI contribution recordTyped roles, activity, delegation, inputs, outputs, model, timestamp, and review disposition.

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://example.org/schemas/contribution-record.schema.json",
  "title": "Contribution Record",
  "type": "object",
  "required": ["id", "target_id", "agent", "roles", "activity", "timestamp", "review_status"],
  "properties": {
    "id": {"type": "string"},
    "target_id": {"type": "string"},
    "agent": {
      "type": "object",
      "required": ["name", "type"],
      "properties": {
        "name": {"type": "string"},
        "type": {"enum": ["Person", "Organization", "SoftwareAgent"]},
        "provider": {"type": ["string", "null"]},
        "model": {"type": ["string", "null"]}
      }
    },
    "roles": {"type": "array", "items": {"type": "string"}},
    "activity": {"type": "string"},
    "acted_on_behalf_of": {"type": ["string", "null"]},
    "inputs": {"type": "array", "items": {"type": "string"}},
    "outputs": {"type": "array", "items": {"type": "string"}},
    "timestamp": {"type": "string", "format": "date-time"},
    "review_status": {"enum": ["unreviewed", "accepted", "accepted-with-edits", "rejected"]},
    "reviewed_by": {"type": ["string", "null"]},
    "notes": {"type": ["string", "null"]}
  }
}

Enterprise research OKF profileResearch governance fields layered over the minimal OKF frontmatter contract.

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://example.org/schemas/enterprise-research-okf-profile.schema.json",
  "title": "Enterprise Research OKF Concept Profile",
  "type": "object",
  "required": ["type", "title", "timestamp", "status", "owner", "provenance"],
  "properties": {
    "type": {"type": "string"},
    "title": {"type": "string"},
    "description": {"type": ["string", "null"]},
    "resource": {"type": ["string", "null"]},
    "tags": {"type": "array", "items": {"type": "string"}},
    "timestamp": {"type": "string", "format": "date-time"},
    "status": {"enum": ["hypothesis", "supported", "validated", "active", "review-required", "superseded", "retired"]},
    "owner": {"type": "string"},
    "confidence": {"enum": ["high", "medium", "low", "not-applicable"]},
    "relationships": {"type": "object", "additionalProperties": {"type": "array", "items": {"type": "string"}}},
    "provenance": {
      "type": "object",
      "required": ["directed_by", "generated_by", "review_status"],
      "properties": {
        "directed_by": {"type": "string"},
        "generated_by": {"type": "array", "items": {"type": "string"}},
        "review_status": {"type": "string"},
        "source_ids": {"type": "array", "items": {"type": "string"}}
      }
    }
  },
  "additionalProperties": true
}

Dataset summary

{
  "metadata": {
    "title": "Enterprise Communication Research System",
    "version": "1.1.0",
    "status": "Living research synthesis",
    "generated": "2026-07-17",
    "description": "A canonical research and communication system for technical documentation, academic research, architecture communication, personas, persuasion, business value, trade-offs, LLM retrieval, and future-agent stewardship.",
    "method": "Synthesis of standards, university guidance, peer-reviewed and emerging research, and first-party enterprise practices.",
    "evidence_policy": "Evidence classes, source intent, scope, limitations, and company-practice boundaries should remain explicit.",
    "last_updated": "2026-07-17",
    "canonical_format": "OKF-compatible Markdown/YAML bundle compiled into JSON, JSON-LD, and a self-contained SPA.",
    "authorship_policy": "Human ownership and accountability remain explicit; AI systems are recorded as software-agent contributors with bounded roles, inputs, outputs, and review status."
  },
  "counts": {
    "principles": 12,
    "domains": 13,
    "genres": 28,
    "findings": 80,
    "playbooks": 18,
    "recommendations": 38,
    "references": 90,
    "schemas": 9
  },
  "authorship": {
    "research_owner": {
      "name": "Jesse Graupmann",
      "type": "Person",
      "roles": [
        "research director",
        "author",
        "editor",
        "accountable owner"
      ],
      "responsibility": "Defines goals, directs research, evaluates conclusions, approves publication, and retains accountability for the work."
    },
    "ai_assistance": {
      "provider": "OpenAI",
      "product": "ChatGPT",
      "model": "GPT-5.6 Thinking",
      "type": "prov:SoftwareAgent",
      "roles": [
        "research discovery",
        "source analysis",
        "synthesis",
        "drafting",
        "schema design",
        "HTML and JSON generation",
        "validation"
      ],
      "relationship": "Operated under the direction of Jesse Graupmann as a research and production tool.",
      "accountability": "The software agent is credited for its contribution but is not the accountable author or owner."
    },
    "source_attribution_rule": "The creator or publisher of every underlying source remains independently attributed. Human authorship of the synthesis does not replace source authorship.",
    "review_status": "Human-directed and reviewable; each future material revision should record the responsible person, software agent, source inputs, accepted changes, and rejected changes."
  },
  "okf_profile": {
    "name": "Enterprise Research Profile for OKF",
    "version": "1.0.0",
    "concept_types": [
      "Research Source",
      "Source Annotation",
      "Research Claim",
      "Evidence",
      "Warrant",
      "Assumption",
      "Concept",
      "Interpretation",
      "Persona",
      "Situation",
      "Audience Impact Contract",
      "Argument",
      "Counterargument",
      "Trade-off Analysis",
      "Recommendation",
      "Decision",
      "Challenge",
      "Procedure",
      "Metric",
      "Playbook",
      "Output Manifest",
      "Contribution Record",
      "Provenance Activity",
      "Generated Artifact",
      "Schema Profile",
      "Validation Result"
    ]
  },
  "object_types": [
    "source",
    "annotation",
    "claim",
    "evidence",
    "warrant",
    "assumption",
    "concept",
    "interpretation",
    "recommendation",
    "decision",
    "counterargument",
    "limitation",
    "procedure",
    "metric",
    "persona",
    "situation",
    "impact-contract",
    "challenge",
    "output-manifest",
    "provenance-activity",
    "contribution-record",
    "generated-artifact",
    "schema-profile",
    "validation-result"
  ]
}