TLDR; Before your team goes any further with AI, leadership should settle four decisions: who owns AI choices, what data stays out, which tasks AI may help with, and what happens when something goes wrong. Those four answers will turn quiet, scattered experiments into practice your organization can stand behind.
Is my team already using AI?
Almost certainly. The development director drafts appeal letters with ChatGPT. The program coordinator asks it to tidy the volunteer schedule. Someone in operations summarized last quarter's grant report in thirty seconds using Claude and told no one.
They kept quiet for good reasons. Nobody told them whether it was allowed, and nobody wants to admit a shortcut in an organization that prizes care. The use is real, but the permission is missing.
What is the Judgment Gap?
The Judgment Gap is the distance between what AI can do and what your organization has actually decided about AI. The tools change weekly. Your decisions, if any are on record, may date from a board meeting two years ago. Every week the gap stays open, staff fill it with their own judgment, one browser tab at a time.
Good people make reasonable guesses. Twelve reasonable guesses still add up to twelve different informal policies.
What does leadership need to decide first?
Who owns AI decisions. Name one person, or a small group, to answer questions and approve new tools. Ownership turns "Can I use this?" from a whispered worry into a two-minute conversation.
What data never goes in. For a food bank, the list writes itself: client names, addresses, household income, eligibility details, and anything a client shared in confidence. Write it down plainly. Your staff want this line drawn; most of them have wondered where it sits.
Which tasks AI may help with. First drafts, summaries of public documents, and brainstorming make good starting points. Decisions about people, services, and eligibility stay with people. A short list of approved tasks gives your team room to work and spares them the guessing.
What happens when something goes wrong. Someone will paste the wrong spreadsheet. Decide now whom they tell, how fast, and what happens next. A clear path for mistakes gets mistakes reported, and reported mistakes are the ones you can fix.
How do I start without a months-long policy project?
Start with a conversation. Gather your team, say plainly that you know AI is in use, and ask what they use it for. Frame it as an amnesty. You will learn more in an hour than a survey could tell you in a month.
Then write one page of interim guidance that answers the four decisions above. Date it, share it, and promise a revision. A page your staff can read today serves them better than a polished policy that arrives next fiscal year.
The full policy, the one your board signs, can follow, and by then you'll know what truly belongs in it.
Frequently asked questions
What should an executive director decide about AI before staff use it?
Decide who owns AI choices, what data stays out, which tasks AI may help with, and how staff report mistakes. Those four decisions close most of the Judgment Gap.
Should we ban AI until we have a policy?
A ban usually moves AI use out of sight. Interim guidance keeps it visible while you build the full policy.
Does a small nonprofit need an AI policy?
Yes, and small teams often benefit most. With fewer people, one person's guess becomes the organization's practice faster.
Where does your organization stand?
The Lansby AI Readiness Assessment shows where your Judgment Gap sits and which decision to make first. It's free and takes about 15 minutes. [Take the AI Readiness Assessment →]
Did we use AI for this blog post? Yep! Claude Opus 5.5 is helping us outline and draft posts that help us reply to specific audience questions with our expert-led answers. We input our expertise, revise, copyedit, and publish.
