The Engine Already Knew

The first AI Readiness report worked.

Technically.

It produced a score.

It generated recommendations.

It looked like a finished report.

But something felt missing.

My first instinct was to make the engine smarter.

Instead, we decided to look more closely at what it already knew.

What changed

The engine could already identify the business.

It understood the offer.

It recognised the audience.

It had found trust signals.

It could point to the evidence behind every conclusion.

The report simply wasn't exposing most of that understanding.

So we stopped changing the engine.

We started improving the window into it.

Why it mattered

That discovery changed the roadmap.

Our first instinct had been to improve the Understanding Engine.

It turned out the engine wasn't the problem.

The presentation layer was.

We weren't missing intelligence.

We were missing visibility.

Those are very different engineering problems.

Looking back, that was another reminder that explainable systems don't become more useful simply because they know more.

They become more useful when people can clearly see what they already know.

What I learned

It's easy to assume software needs more intelligence.

Sometimes the biggest improvement comes from exposing the work it has already done.

Spark

Inspired by:

• Internal sprint review