The Tens: KM and Enterprise AI―10 Actions KM Leaders Should Take to Ensure More Robust Enterprise AI

The Tens: KM and Enterprise AI―10 Actions KM Leaders Should Take to Ensure More Robust Enterprise AI

As a follow-up to my post, Knowledge Management and AI: Revisiting the Need to Apply KM Practices and Technology to Enterprise AI Success, which covers the core points of my APQC 2025 talk, I thought it would be useful to create a list of actions knowledge managers should take to start the engagement with their AI development and deployment counterparts. It’s time for KM and enterprise AI teams to talk again…or at least talk more.

(Coverage image by Google ImageFX from a prompt written by the author.)

Knowledge Management and AI: Revisiting the Need to Apply KM Practices and Technology to Enterprise AI Success

KM and Enterprise AI: Knowledge management practitioners should take the following actions to support responsible, strategic AI implementation within their organizations:

1. Establish Model Metadata Standards. Define and enforce metadata schemas for all AI models, including version, training data, fine-tuning details, and intended use cases.
2. Catalog Prompts as Knowledge Assets. Treat prompts as reusable artifacts. Create version-controlled libraries with annotations, tags, and documented performance insights.
3. Create Guardrail Governance Frameworks. Document the intent behind each guardrail. Track changes, run behavioral audits, and support user education to ensure transparency.
4. Map Context Configurations to Business Goals. Review and document context window settings. Ensure they reflect relevant knowledge priorities and are adapted as organizational needs evolve.
5. Audit and Align Knowledge Graphs and RAG Sources. Vet sources for accuracy, bias, and recency. Maintain governance over both curated knowledge structures and dynamically retrieved content.
6. Implement Model Lifecycle Management. Track usage patterns, test responses across versions, and archive model-related decisions to preserve institutional knowledge across transitions.
7. Develop Frameworks for Communicating Uncertainty. Help AI systems disclose the limits of their knowledge. Educate users to interpret generative output as a starting point, not an endpoint.
8. Facilitate Cross-Functional Collaboration. Embed KM into AI development and deployment teams to ensure that epistemic decisions are made with context, governance, and business alignment in mind.
9. Support Prompt and Guardrail Communities of Practice. Create space for iterative improvement by bringing together those designing, testing, and refining prompts and constraints across use cases. Agent orchestration teams should also form communities of practice to ensure the efficiency and consistency of agent-based AI solutions.
10. Monitor Emerging Knowledge Gaps. Establish feedback loops (or event triggers for new products, marketing campaigns or other events) to detect when models fall behind reality. Create protocols to accelerate the inclusion of new, validated knowledge.

These actions elevate KM to an essential role in AI success by ensuring that AI systems are understandable, auditable, usable, accurate, robust and aligned with organizational strategy.

For more “tens” from Serious Insights, click here. For more serious insights on AI, click here.

Did you like this post on KM and enterprise AI? If so, like, comment and share the post.