Almost every nonprofit uses AI. Few have an AI playbook.

TL;DR

You're not behind, and you're not alone

In a 2026 survey of 917 nonprofit staff and leaders by NTEN and The Bridgespan Group, 98 percent said their organization uses AI in some capacity. About 39 percent of executives said written guidance on responsible AI use is in place, and another 38 percent said it is still in development. Only about 22 percent said a plan for managing AI-related risks is in place.

Using AI is not the same as being changed by it. In a separate survey of 346 nonprofits at the end of 2025, 92 percent used AI in some way, while only 7 percent said it had produced major improvement in their ability to achieve their mission. The numbers will keep moving. The shape has stayed the same: wide use, thin structure.

There is also a gap inside organizations. Nonprofit Quarterly's summary of the NTEN report notes that more than half of respondents reported AI use outside organizational guidance, and that executives were more comfortable with AI than staff were. When people are using a tool nobody has talked about, the problem is not the people. It is the missing conversation.

Why using AI is not the same as adopting it

The tools change every few months. What decides whether AI sticks inside an organization is whether people trust it, have a voice in how it is used, and know where the lines are. That is why a policy handed down from the top rarely changes behavior, and why a tool rolled out without conversation rarely lasts.

The five steps below come from the AI adoption boot camp taught by AI Coachella Valley (AICV), a local AI organization. The boot camp is facilitated, and it is built to help an organization develop its own capacity to adopt AI responsibly, not to hand it a set of automated workflows. Think of them as a ladder. Each rung holds the next one up, and skipping a rung is how adoption stalls.

The ladder: five steps from first use to lasting practice

1. Commit

Leadership makes a real commitment to AI adoption, including time, resources, and visible support.

What it looks like: a named executive sponsor, protected hours on someone's calendar, and a modest line in the budget, even if it is small. People can tell the difference between a priority and a hope.

Where it goes wrong: AI becomes a side project that disappears during the first busy season.

Your first move: put two hours a week on a named person's calendar for the next ninety days.

2. Represent

Create a cross-functional AI Council, so adoption is not owned only by IT or leadership.

What it looks like: three to five people across programs, development, and operations, plus a board liaison. In a small nonprofit the council might simply be the people who are already using AI every day.

Where it goes wrong: the council is made up only of managers, so the playbook ends up describing a nonprofit that does not exist. The people closest to the work need to be in the room.

Your first move: ask who is already using AI, and invite them.

3. Listen

Give employees a safe place to surface concerns, ideas, opportunities, and the work they actually want AI to help with.

What it looks like: a short anonymous form, or twenty minutes in a monthly meeting built around three questions. What did you try? What worked? What worried you?

Where it goes wrong: leaders ask after the decisions are made, or react badly to honest answers. People then stop telling you what they are doing, and use goes underground.

Your first move: ask every person one question: what is one task you wish AI could take off your plate, and what is one worry you have?

4. Create the AI Playbook

Develop a living organizational guide covering approved tools, use cases, guardrails, workflows, lessons learned, and evolving best practices.

What it looks like: it starts as one page and grows. It answers a handful of plain questions. Which tools are approved? What information must never go into them, such as donor records, client details, or anything covered by a privacy promise? When do you tell people AI helped? Who reviews AI output before it goes out? Who owns the playbook? How does a staff member raise a concern?

Where it goes wrong: it becomes a twelve-page policy no one reads, or a template copied from somewhere else that does not match the tools and data you actually have.

Your first move: write the one page. Imperfect and in use beats complete and unread.

5. Evolve

Keep experimenting, training, measuring, sharing what works, and updating the playbook as the technology changes.

What it looks like: a review on the calendar every quarter, a running log of what you learned, and training in the formats people say they want. In the NTEN survey, staff preferred interactive tools and exercises, and live workshops, far more than passive formats. Measure something simple, such as time saved on a task or errors caught before they went out.

Where it goes wrong: the playbook is written once and never opened again, while the tools change underneath it.

Your first move: put the first review on the calendar, ninety days from now, before you do anything else.

Values the playbook can stand on

A nonprofit runs on trust, and trust is the first thing careless AI use puts at risk. AICV's Responsible AI Pledge offers seven plain-language principles that translate directly into playbook language: People Lead, AI Follows. Be Honest About What's AI. Guard What's Private. Watch for What's Unfair. Think About Real People. Share What You Learn. Stay Human.

The sector has begun to write down similar expectations for its own work. The Fundraising.AI framework, developed by a collaborative of fundraising professionals and platforms, calls for ongoing human review of any AI system that makes critical decisions, clear disclosure when AI shapes donor interactions, and care with donor data. Your playbook does not need to be as long as a framework. It needs to say, in your own words, where your organization stands.

Where to start, close to home

AICV teaches the boot camp described here, and also runs hands-on AI workshops across the Coachella Valley. To learn more, visit aicoachellavalley.org.

Whatever you do next, start with the people. The tools will change. A team that feels included, trusts its guardrails, and knows how to keep learning will adapt to whatever comes next.

The boot camp described here is taught by AI Coachella Valley (AICV), a fiscally sponsored project of Desert Community Foundation, the founder of CV Giving Day.