Notes for owners · July 2026

What should a business owner use AI for first?

The wrong first AI project is a moonshot. A chatbot on the website. A “strategy assistant.” Something impressive in a demo and useless on a Tuesday. Six weeks later the tool is abandoned and the owner concludes AI isn’t ready.

The right first project is boring, and that’s exactly why it works. It’s the recurring document that eats the most of your month. The monthly client report. The estimate cycle. The quarterly deck. You drafted it last month. You’ll draft it next month. It has a format, a standard, and a deadline, which means success is measurable and failure is obvious.

The three tests

Run any candidate through the filter we use on every build:

• Does it repeat? If you’ve done it ten times, there’s a pattern, and a pattern can be trained.
• Could you brief a great assistant on it? If you can explain how it should be done, it can be built.
• Would you correct an employee who did it wrong? Then you already have a standard the system can follow.

Everything that passes all three is a candidate. Then pick by hours: whichever one costs you the most owner time wins. (If you haven’t added those hours up, the math is usually worse than owners expect.)

Why the first project matters so much

The first project sets the pattern for everything after it. When the monthly report comes off your plate in month one, drafted in your voice, at a standard you’d sign, two things happen. You get hours back immediately, and you learn to trust the system on work where trust can be verified. That trust is what lets the second and third projects move faster.

Start with the moonshot instead and neither happens. There’s no baseline to measure against, no standard to test, and no Tuesday where you feel the difference.

What this looks like in practice

For a fractional CFO, it’s the quarterly deck. For a home service company, it’s the estimate-and-follow-up cycle. For an agency, it’s monthly reporting. The industry changes the document; the logic doesn’t: recurring, briefable, held to a standard.

If you want the fuller version of this decision framework, the dream-employee test is the way we teach owners to run it, and the FAQ covers the questions that usually come next.