Picture your most experienced project executive. Thirty years in. Knows where every body is buried on every project type. Knows which subcontractors to trust for what, which details fail in your climate, which client communication styles to match.

Now picture them walking out the door.

How much of that knowledge goes with them?

Most of it. Because most AEC firms have no systematic way to capture what their best people know. It lives in email threads, in meeting notes that receive no sustained attention, in tribal memory that transfers through proximity and conversation. When those people leave (to retirement, to a competitor, to burnout) the knowledge leaves too.

The custom GPT answer isn’t the point

There’s a lot of energy right now around building AI tools: custom GPTs, Azure AI projects, internal chatbots. The tools are real and they’re useful. But the tool is not the insight.

The insight is that for the first time, it’s affordable and practical to build searchable, queryable institutional memory. To take what your senior people know (the decisions they’ve made, the problems they’ve solved, the lessons they’ve learned) and make it accessible to everyone in the firm.

A proposal bot that knows your firm’s past projects, your best case studies, your standard scopes and fees, your differentiators, that’s not a gimmick. That’s a competitive advantage on every pursuit. An onboarding tool that can answer a junior employee’s questions the way a senior colleague would, that’s not replacing mentorship, it’s extending it.

What owners should notice

You’re on the other side of this problem too.

When you work with a firm that has captured its institutional knowledge well, you get more consistent service. The PM on your project isn’t reinventing the wheel on questions your firm has answered twenty times before. The proposal they submitted actually reflects what they know, not just what they could pull together in a week.

When you work with a firm that hasn’t, you pay for the learning curve on your dime. You get the junior team that hasn’t done your building type before. You get the change orders that stem from not knowing what they didn’t know.

Ask your partners how they transfer knowledge across projects and across people. Ask what happens when a key team member rolls off. The answer tells you something real about organizational quality.

The right starting point

Here’s what firms get wrong: they start with the tool question instead of the knowledge question.

Which AI platform? Azure or OpenAI? Build or buy? These are secondary. The primary question is: what knowledge do we actually need to preserve, and where does it currently live?

Start there. Map the knowledge. Find the pockets where hard-won experience lives in people’s heads rather than in systems. Then figure out the most practical way to start moving it.

The technical part is easier than it’s ever been. The organizational part (getting people to contribute, creating the habit of capturing rather than just doing) that’s where the real work is.

Firms that solve that problem compound. Their best people leave behind something that keeps working. Their junior staff get smarter faster. Their proposals get sharper. Their projects get better.

The firms that don’t solve it just keep starting over.