AI-POWERED KNOWLEDGE CREATION
Draw out tacit expertise through generative AI dialogue—and bring it
into shared knowledge, collaboration and learning.
AISECI provides a framework for putting experience into words and combining it with organizational information. Its 12 capability areas define a target architecture, including development directions. We establish which capabilities can be deployed and which require further development for each engagement.
Organizations accumulate experience, information, and practical know-how. Making these resources useful across teams requires context, review, and opportunities to apply them.
Experienced staff develop judgment, practical workarounds, and lessons from customer interactions and past failures.
The knowledge that matters most in practice is often the hardest to put into words, so it stays within people’s experience. That makes it difficult to share and pass on, and it risks being lost when people change roles or leave the company.
Relevant information can be spread across documents, email, chat, meeting records, databases, and business systems.
Even when there is plenty of information, if each source is siloed it becomes hard not only to find what you need but also to combine different pieces of information into new insight.
Manuals and knowledge bases are more useful when teams can find and apply them in their work.
Without feedback and review, documented guidance can drift away from current operating conditions.
The SECI model describes organizational knowledge creation through interaction between tacit and explicit knowledge. Ikujiro Nonaka presented the theory in 1994; Nonaka and Hirotaka Takeuchi developed it further in The Knowledge-Creating Company (1995).
The model connects shared experience, articulation of insights, integration of explicit knowledge, and learning through practice. AISECI uses these processes as a design framework; the model itself does not establish the effectiveness of an AI implementation.
Share experience and pass knowledge from person to person.
Put experience and judgment into words and turn them into knowledge others can share.
Combine multiple pieces of knowledge into more systematic knowledge.
Put knowledge into practice and make it part of your own judgment and skills.
In the SECI model, the four processes interact over time rather than forming a one-time sequence.
Socialization▶Externalization▶Combination▶Internalization
Repeated dialogue, reflection, and practice can contribute to new insights.
The knowledge spiral describes how insights can develop from individual experience into shared team and organizational knowledge. This depends on participation, context, and opportunities to apply what is learned.
Source: Ikujiro Nonaka, Hirotaka Takeuchi, The Knowledge-Creating Company: How Japanese Companies Create the Dynamics of Innovation, Oxford University Press, 1995 (Japanese edition: Toyo Keizai, 1996)
Earlier theory: Ikujiro Nonaka, A Dynamic Theory of Organizational Knowledge Creation, Organization Science 5(1), 1994. The diagrams on this page are conceptual illustrations and AISECI’s capability mapping is a design interpretation.
Human experience, dialogue, and learning remain central to knowledge creation. As organizations grow, collecting and reviewing experience, connecting information, and supporting learning can become harder to coordinate.
AISECI’s target architecture combines human expertise with AI-assisted information processing. The roles below describe the intended approach, with people responsible for validating interpretations, approving guidance, and assessing its use.
Digital tools support the storage and organization of explicit knowledge. Generative AI can also assist with interview prompts, summaries, and comparisons that help people reflect on experience.
Experts can use dialogue to explain observations, decisions, and rules of thumb. AI can help draft these explanations for review, but cannot directly access unspoken experience or fully capture a person’s tacit knowledge.
Shared work, observation, and dialogue support learning between people. AI-assisted prompts and practice conversations are intended to support these human activities; a simulated dialogue does not replace direct experience with an expert.
Human-led / AI-assistedThe intended approach uses interview prompts and conversation analysis to help experts articulate decision criteria and rules of thumb. AI-generated interpretations and drafts require expert review before becoming shared guidance.
Human knowledge / AI elicitationSearch, classification, and relationship extraction can help organize explicit knowledge from approved sources. The target design retains relevant source context and access permissions, with proposed changes reviewed before publication.
AI-assisted integration / Human reviewAI-assisted exercises, feedback, and contextual guidance are intended to support learning. People develop skills through practice; training owners assess content quality, and users remain responsible for workplace decisions.
AI-assisted learning / Human practiceThe 12 areas below describe AISECI’s target architecture, not a list of 12 released features. For each engagement, we confirm available components, feasibility, required development, and integration scope before agreeing on delivery.
Select a number to explore its intended role.
Intended to support practice dialogues grounded in approved expert interviews and documents. Responses would be AI-generated approximations for learning, not statements by the expert or a complete reproduction of their judgment. Source scope and expert review would be agreed during design.
Intended to identify candidate insights and stated reasons for decisions in approved meeting, sales-call, and interview records. Participants or subject-matter experts would review interpretations; recording, use, and access requirements would be agreed.
Intended to simulate customer or workplace roles for dialogue practice using reviewed scenarios. Simulated responses would support rehearsal, without representing real customers or replacing shared workplace experience.
Intended to draft meeting minutes, procedures, FAQs, and standard operating procedures from approved conversations and work records. Responsible staff would validate accuracy, completeness, and applicable rules before publication.
Intended to guide interviews that help experts explain decision criteria and rules of thumb. Proposed interpretations would be checked with the expert; information not articulated or recorded cannot be assumed.
Intended to compare documented experiences and propose recurring patterns for review. Similarities would not, by themselves, establish causation or justify a general business rule.
Intended to retrieve relevant passages from connected, approved sources to support source-backed responses. Coverage, update frequency, user permissions, and response quality would be established and evaluated for the deployment.
Intended to represent defined entities and relationships with links to supporting sources. Domain definitions and extracted relationships would require validation; a graph does not itself establish that a statement is true.
Intended to select guidance using agreed role and task context within the user’s access permissions. Relevance would depend on the available context and the quality of the underlying knowledge.
Intended to adapt reviewed learning content using agreed signals such as role, stated experience, and exercise results. Adaptation rules and feedback quality would need evaluation; these signals do not directly measure a person’s understanding.
Intended to generate practice exercises from approved workplace scenarios, with reviewed evaluation criteria and feedback. Practice performance would not be treated as proof of competence in real operations.
Intended to surface relevant guidance and proposed next steps within an agreed workflow. System integration, latency, access, approval requirements, and escalation paths would determine its operating scope.
This example illustrates how selected capability areas could support sales knowledge sharing. It is not a report of a completed deployment or measured results. Availability, development needs, data access, and evaluation would be confirmed for the engagement.
Sales conversations can provide evidence of customer responses, questions, proposal choices, and stated reasons for decisions. Some experience remains unspoken and needs follow-up with the people involved.
In this proposed workflow, Conversation Mining would identify candidate insights for team review. Expert Persona Simulation would support practice dialogues based on approved material, alongside direct discussion with experienced salespeople.
The team would review approved call records and deal histories, using AI-assisted drafts to identify observations and questions for follow-up.
Knowledge Extraction would help participants explain decisions and possible contributing factors. Automated Documentation would draft playbooks, scripts, and FAQs for review. Deal outcomes alone would not establish why a sale succeeded or failed.
Reviewed sales guidance could be connected to relevant CRM records, prior proposals, and product information through agreed integrations and access controls.
Enterprise RAG would retrieve supporting information, while the Knowledge Graph would link defined customers, needs, products, and cases. Users would retain access only to permitted information.
Applying reviewed guidance and reflecting on outcomes can help salespeople develop their skills.
Scenario Exercises would support preparation for customer conversations. An On-the-Job AI Assistant could surface guidance in the agreed workflow; its timing and integration would need validation. Salespeople would review communications and retain responsibility for commercial decisions.
Participants would discuss outcomes and propose updates for review before incorporating them into shared guidance.
Review new experience and share lessons with the team
Help teams document experience, decision criteria, and lessons from past work as reviewed, reusable guidance.
The aim is to make recorded expertise easier to share during onboarding and handovers, while retaining access to experienced staff.
Connect relevant sources and make reviewed guidance easier to find within agreed workflows. Usefulness depends on data coverage, integration, and user adoption.
Use feedback and documented outcomes to propose knowledge updates. Review and evaluation are needed to keep guidance relevant and assess whether it improves work.
Begin with one process where knowledge sharing is a clear need. Assess available components and required development, evaluate a defined scope with representative data, and expand only where results and operating conditions support it.
Analyze the target work to clarify which knowledge matters, who holds it and where the flow of knowledge stalls.
Identify approved documents, records, and expert interviews. Agree on permissions, recording requirements, data handling, and review responsibilities before collection or integration.
Select relevant areas from the target architecture and confirm component availability, feasibility, and development needs.
Agree on integrations, acceptance or exploratory evaluation criteria, and human review controls before building and testing the selected workflow.
Review usage, feedback, and business outcomes against agreed objectives. Validate proposed knowledge or system changes before release.
Discuss a focused evaluation of AISECI for your knowledge-sharing needs. We will assess available components, required development, and the scope appropriate to your workflow.
