A workbook for nonprofit boards, directors and staff deciding whether, and how, to use an AI tool. Pick one tool or one planned use, then answer 50 questions about it before the first seat is bought or the first client record goes in.

It opens with a real case: in 2023 a national eating disorders helpline moved to a chatbot that told users to count calories, and had to take it offline. Every question in the workbook is one somebody should have asked first.

50 questions in five sections. Mission and people; data, privacy and consent; bias and fairness; accountability and governance; staff and culture. Each question asks one thing and has Y, P and N boxes to tick.

Eight STOP questions. These are marked in red. If the answer to any of them isn't yes, the tool doesn't launch yet.

Ten questions for the board. Picked from the 50, to bring to the next board meeting before anyone signs a contract.

Return on mission. A simple chart for sorting what to pilot, fund anyway, keep to back-office work, or stop.

The next 90 days. What to do this week, by day 30, 60 and 90, ending in a one-page AI use policy and a measured pilot.

Gap log. A printable page to log every gap with an owner and a date.

Which state privacy laws cover nonprofits (California mostly exempts them, Colorado doesn't), what changed in Colorado's AI law, what to ask a vendor about training on your data, and a free framework for bias testing. Laws and vendor terms checked September 17, 2026.

A full-color screen edition for reading and presenting, and a printer-friendly edition for working through it on paper with your team. 19 pages each.

Once your team has answered the questions, I can help turn the answers into an AI use policy and a pilot plan.