A benchmark study released this month surveyed 346 nonprofits about how they actually use AI day to day, not the five-year vision from a strategic plan, and the results describe almost every organization I've worked with in Cochise County. Ninety-two percent of nonprofits are now using AI in some form. Only 7 percent say it has made a major difference to their ability to achieve their mission. That gap between "everyone's using it" and "almost nobody sees results" is worth sitting with, because it's not really a story about the technology. It's a story about what happens after the free trial ends.
The Numbers Behind "Everyone's Already Using It"
The 92 percent figure, from Virtuous's 2026 Nonprofit AI Adoption Report, tracks closely with what Google found when it surveyed more than 6,000 organizations enrolled in Google for Nonprofits: 86 percent agreed AI has the potential to make their work more effective, but on average less than half, 49 percent, said their organization actually uses it regularly. Read together, the two surveys tell the same story from opposite ends. Nearly everyone has tried AI. Roughly half use it with any consistency. And only a sliver report it changing what they're capable of as an organization. One fundraising leader quoted in the Virtuous report put it plainly: "We're definitely faster. But if you asked me if we've fundamentally changed what we're capable of as an organization, I honestly don't think so. We're just doing the same things more efficiently."
Adoption Without Transformation
Virtuous calls this the efficiency plateau, and it's an accurate name. Most organizations using AI are drafting emails faster, brainstorming appeal language, or summarizing a grant report in a fraction of the time it used to take. Those are real gains, especially for a two-person development office, but they're incremental. They speed up an existing task rather than opening up a new one. The 7 percent seeing major strategic impact aren't doing something more advanced with the models themselves; Claude, ChatGPT, and Copilot work the same for them as for everyone else. What's different is what happens around the tool.
Why the Barriers Change as You Grow
The survey breaks the obstacles out by how far along an organization is, and the shift is telling. Nonprofits that haven't started yet cite lack of training (48 percent), not knowing where to begin (44 percent), and doubts about their own capability (44 percent). Once an organization is already using AI daily, those concerns are almost entirely replaced by a different set: time and capacity constraints (31 percent), privacy and security worries (32 percent), and staff skepticism built up from a bad first experience (19 percent). In other words, getting started and staying disciplined are two separate problems, and most of the advice aimed at nonprofits only addresses the first one.
The Governance Gap
Here's the number that should concern any board member reading this: 47 percent of nonprofits using AI have no governance policy at all, and 81 percent are using it ad hoc, with no documentation of what actually works or who's tried what. UST's own 2026 review of the sector found nearly the same thing, describing it as a widening gap between organizations that treat AI as infrastructure and those still treating it as a personal productivity hack. When a staff member figures out a prompt that saves an hour on grant reporting and then leaves six months later, that knowledge walks out the door with them. Multiple people on the same team end up solving the same problem in isolation, week after week, because nothing was ever written down.
What Separates the 7%
The organizations reporting real transformation aren't spending more money or hiring AI specialists. Virtuous found they've made a small number of foundational decisions that take days, not months, to put in place: a shared document capturing prompts and approaches that actually work, a single-page acceptable use policy agreed on in one meeting, and some basic way of tracking whether a given use of AI is actually producing better outcomes. None of that requires new software. It requires someone deciding it's worth doing before the next new tool shows up.
Why This Hits Small Organizations Hardest
UST's report flags a real risk here for the smaller nonprofits and schools that make up most of Cochise County's sector: larger organizations with dedicated staff and bigger budgets are pulling ahead in how strategically they use AI, while smaller organizations, already stretched thin, stay stuck in the ad hoc, one-person-figured-it-out phase. Left unaddressed, that gap doesn't stay flat. It compounds, because the organizations with more capacity keep building on what they've already documented while everyone else starts over every time someone new picks up the tool.
Where That Leaves You
The good news is that closing this gap doesn't take a large budget, and you may already have a head start. We built a free AI acceptable use policy template and a nonprofit prompt library for exactly this reason, because the first two items on Virtuous's list of what separates the 7 percent, a written policy and a shared set of proven prompts, are things most organizations can put in place in an afternoon rather than a fiscal quarter. If your organization has been using AI for a while and it still feels like everyone's improvising on their own, that's a solvable problem, and a short conversation is usually enough to find the starting point. Reach out through the contact form and let's talk through what that would look like for your team.
