KM + AI Readiness: What the Machine Can’t See

[MaryG]

I’ve been working on a KM + AI Readiness Assessment, and one of the things I keep thinking about is how it fits alongside the new Knowledge Center optimization capabilities.

Article Optimization and Knowledge Health Score are incredibly useful. They can continuously look at the content itself and surface things like duplicates, broken links, accessibility issues, gaps, title relevance, expiration, and other indicators of content health.

We need that. Especially now that the same knowledge is feeding search and AI.

But I don’t think we can let the machine be the entire assessment.

There are things it simply can’t tell us.

Who really owns Knowledge Management?

Is governance documented, and does anyone actually follow it?

Does the taxonomy make sense because someone designed it intentionally, or did it just grow that way?

Are articles being reviewed because there is a sustainable lifecycle process, or because somebody periodically notices the backlog?

Are KM metrics tied to actual business outcomes?

Does the organization understand what it wants AI to accomplish?

And maybe most importantly, does what people believe is happening match what is actually happening in the platform?

That’s why I’m starting to see the concept of a KM + AI Readiness Assessment as two things working together.

The platform can give us objective signals about content health. The consultant can add the organizational, operational, and platform context around those signals.

For the initial assessment, I think that human part matters enough that it shouldn't be fully automated. As a consultant, I want the conversation. I want to hear where answers don’t line up, ask the next question, validate what I can, and establish a baseline the customer understands and trusts.

After that, though, I see a very different role for AI.

Once we have a validated baseline, an agent could help the customer reassess periodically, compare results, identify changes or regression, surface new gaps, and tell them where they may need to take another look.

So maybe the model becomes:

Platform signals + consultant-led discovery → validated baseline → ongoing agent-supported assessment

The optimizer tells us a lot about the health of the knowledge.

The assessment helps us understand the health of the Knowledge Management program around it.

I think we need both.
#KnowledgeManagement #AIReadiness #KnowledgeStrategy #NowAssist #ServiceNow