Adapt · Essay

AI didn’t expose your tech. It exposed your culture.

Greg Moran · October 2026 · 5 min read

The tools are available to everyone now. Someone on your team sees an opportunity, the technology exists, and trying costs almost nothing. If nothing happens anyway, you don’t have a tech problem. You have a “But First” culture, and AI just made it visible.

Who this is for

Founders and CEOs of $5M+ B2B software companies who have bought the licenses, run the pilots, and still feel the company moves at the same speed as last year. Also for operators who suspect the bottleneck sits in a meeting, not in the stack.

The reasoning

Here is the pattern I keep seeing. A person spots an opportunity. Maybe a support workflow that eats forty hours a week, or a sales research task that nobody enjoys. The tool to fix it exists. A prototype would take an afternoon. The cost of trying is close to zero. And then it stalls.

Not because anyone said no. Because of “But First.” But first, let’s get legal to look at it. But first, let’s wait for the Q1 planning cycle. But first, let’s form a working group. Each one sounds responsible. Added together, they are a decision to do nothing.

Before AI, “But First” was expensive but survivable. Building took months, so a few weeks of process felt proportionate. That math is gone. When the build takes an afternoon and the approval takes a quarter, the approval is now the product. Your process is the slowest thing in the company, and everyone can see it.

The second failure is quieter. Some teams adopt AI and make the old workflow faster. Same steps, same handoffs, same approvals, just quicker. It feels like progress. It isn’t. You sped up a workflow that deserved to be rebuilt. AI either makes you smarter and faster or dumber and slow. There is no in between.

A practical method

1. Pick one workflow everyone complains about. Just one.
2. Give one person the authority to rebuild it end to end with AI, inside a week. No working group.
3. Count the “But Firsts” that show up along the way. Ask of each: does this protect something real, or does it protect someone’s comfort?
4. Measure the rebuild against the old workflow in finished work. Not hours saved, not tool usage. Finished work.
5. Ship it or kill it cleanly. Either way, pick the next workflow and go again.

Tradeoffs and limitations

Some “But Firsts” are real. Regulated data, customer-facing risk, security review. The point is not to skip judgment. It’s to refuse ceremony dressed up as judgment. Be honest about which one you’re looking at. And know that a rebuilt workflow will break things the old process quietly handled; budget for that. Don’t run this play on systems where an error is expensive and invisible.

A useful next step

Write down the last idea your team didn’t try. List every “but first” that stopped it. Cross out the ones that protect nothing. What’s left is your real approval process. Shorten it this week.