Inputs
A defined purpose, participating expert, representative situations, organizational boundaries, consent, and relevant source material.
Protocol and principles
A transferable way to prepare, conduct, review, and apply Knowledge Harvesting.
The emerging protocol brings the method together with defined materials, safeguards, human responsibilities, and review criteria. Independent application is being pursued and tested; it is not yet claimed as complete.
Protocol anatomy
What must travel with the method.
Process
Preparation, structured elicitation, classification, synthesis, expert review, correction, application, and learning from use.
Outputs
Reviewed knowledge resources that preserve context, distinguish evidence from inference, and make reasoning available for an intended use.
Safeguards and operating principles
Extend expert judgment with care.
These commitments set the minimum standard for responsible implementation.
- Respect and voluntary participationExperts understand the purpose, boundaries, intended use, and choices available before participating.
- Attribution and ownershipParticipation does not silently transfer ownership. Attribution, reuse, licensing, and organizational rights are agreed explicitly.
- Context preservationKnowledge retains the situations, limits, assumptions, and uncertainties needed for responsible interpretation.
- Expert reviewParticipants can correct, qualify, withhold, or approve what is represented as their knowledge before release or use.
- Confidentiality by designSensitive information is identified, minimized, protected, and handled under agreed organizational and legal requirements.
- TraceabilityOutputs distinguish source material, expert statements, model synthesis, practitioner interpretation, and unresolved questions.
- Human oversightAI may assist inquiry and organization; people remain responsible for purpose, judgment, release, and consequential use.
- Appropriate escalationA resource states when a user must pause, seek qualified judgment, or return to the expert rather than act from the artifact alone.
- Technology neutralityThe method is technology-neutral, not dependent on one model or vendor, and conforms to the approved environment and policies of the organization using it.
- Continuous improvementUse reveals omissions and ambiguity. Resources and protocols are revised from evidence rather than treated as final because they were published.
Data, AI, and human review
Answer the questions that determine trust before elicitation begins.
Who owns the expert's contribution? What will be recorded and retained? What may enter an AI system? Which environment is approved? Who may review the material? How will sensitive information be protected? Who can authorize an output for use?
AI may help ask questions, organize responses, find patterns, draft structures, and expose gaps. It cannot assume authority. Human participants establish purpose, judge relevance, protect sensitive material, resolve ambiguity, and approve what is represented as expert knowledge.
Public and professional resources
Explain the boundary now; define programs only when they are real.
Public foundation
Method overview, principles, evidence boundaries, development status, introductory resources, and opportunities for careful review.
Professional capability
Advanced instruments, implementation materials, validation tools, facilitation, training, and partner rights may be licensed or supported professionally as they are defined.
No certification system, partner network, or formal licensing program is being announced. Such structures should be described only after their obligations, safeguards, governance, and evidence requirements are settled.
Current development
96 is an implementation surface, not the definition of the method.
96 explores the Knowledge Harvesting questioning architecture at finer resolution: eight kinds of knowledge across twelve information types. Its question architecture, elicitation agents, orchestration, and outputs are being completed, tested, and aligned.
Independent practitioner application, consistent use, and validated output criteria have not yet been established publicly. 96 is therefore presented as active development, not as a finished or independently validated instrument.