Organizations are sitting on a lot of expertise. Knowledge is abundant. Yet we lack the ability to turn that expertise into consistent performance.

Especially now, as organizations move toward skill-based models, which require building capability faster and more consistently across roles.

And we are still using the same approach: more training, more content, more tools (now AI).

Result: overloaded SMEs, stretched L&D teams, and learning not matching real work.

So how can we translate expertise into practice better?

Why don’t we consider:

• stopping the centralization of everything through L&D
• and start enabling SMEs to contribute where learning actually happens (inside the work itself)?

L&D doesn’t disappear in this model; its role becomes more critical.

It focuses on alignment, quality, and structure, instead of becoming a content factory.

This only works if SMEs are not asked to “teach” in the traditional sense (with L&D-designed slides for 40 hours).

They just need a simple way to structure what they already do:

• What does good work look like?
• In which situations does their skill matter?
• What should someone do differently?

From there, learning becomes part of the workflow:

• a quick preparation moment before the task,
• guided application during,
• and short reflection after.

No heavy programs. No extra layers.

Just better use of moments that already matter.

This is also where AI starts to make sense, but only if the foundation is there.

If expertise is not structured, AI simply generates more content and noise.

If it is structured around real scenarios and decisions, AI can extend it:

• reinforcing key moments in real time
• reducing repetition for SMEs
• helping learners navigate situations as they happen

Not as a replacement, but as support.

It’s about making expertise usable in the flow of work.

Because the real constraint is not SMEs.

It’s how the system around them is designed.

***The sculpture is by artist Willy Verginer.