I Hadn't Found the Value of AI. Then I Handed It the Right Problem.

Most of us are using AI in some form by now. But have you actually found the high-level value we were all promised?

I hadn't. I used it as a research tool. A quick cross-check on something I'd written. Useful, but nothing close to the work transformation everyone kept talking about.

Then something changed.

I was part of a large group vetting a complex new system — one that touches our entire organization. It had to handle heavy usage, keep impeccable records, and correctly ingest years of historical data. Customer records, especially, had to be flawless. Or as close to it as we could get.

We ran two or three sessions a day for over a week. Different facilitators. Attendees moving between rooms. Note-takers in every session, plus full transcripts as backup. The volume was enormous.

And leadership wanted a gap analysis and key decisions by day three.

We knew that wasn't realistic. Best case, the following week. More likely two weeks out.

So I tried something. I fed the transcripts from two sessions — the ones that wrapped up a major area of concern — into an AI tool. Within minutes, I had three pages of decisions and gaps. Categorized. Organized. Easy to read.

We weren't done, of course. I still had to cross-check the output against the other documentation from those sessions. But suddenly I had something to put in front of leaders. Something that showed them we were capturing what they'd asked for. A solid first step toward answering whether this new software could actually do what we needed.

I was clear with the team: this was a first draft. An example. We'd still confirm everything. We'd still find mistakes — no more than I would've made doing it the long way, but they'd be there.

What I learned surprised me. AI is a far more powerful tool than I'd given it credit for. It saved me hours, possibly weeks. It can take the first pass so I can focus on the real thinking — and get to steps two and three much faster.

Is it flawless? No. But neither am I. Doing this work alone, I would have made the same mistakes, probably more of them, and taken exponentially longer to deliver.

That's the value I'd been waiting to find. And I think a lot of us are closer to it than we realize — we just haven't handed it the right problem yet.