You advise clients, run a firm, and sit on boards. All three jobs are mostly reading, drafting, and deciding. AI is genuinely good at the first two — and dangerous at the third. Here's how to tell the difference.
See the playbook ↓Plain English. No jargon. Built for people whose time is billed.
Not theory. The specific work you can move off your desk this week.
Most disappointing AI results are a one-line-prompt problem, not a capability problem. Same document, same model, thirty seconds more effort.
Not hypothetical. Lawyers have been sanctioned for filing citations the model invented — and it keeps happening.
The one-line test: before you paste anything — would you be comfortable if this exact text showed up in a discovery request against you? If not, strip the identifying details. The analysis is nearly always just as good.
The goal isn't to learn AI. It's to end the month with real work permanently off your plate.
A contract you've already reviewed. A packet you've already read. You're calibrating trust, not saving time yet.
Your three most repeated tasks, written out properly once and saved. You'll reuse them hundreds of times.
One page for the firm — and share it with your boards. Most nonprofits have no policy at all and will thank you.
Intake summaries, board briefs, billing narratives — pick one, make AI the permanent first step, you the reviewer.
Take the board packet or contract that lands tomorrow. Run it through a proper four-part request. Compare what comes back against your own read. That one comparison teaches you more than any course.
Back to the playbook ↑