AI tools are multiplying faster than teams can evaluate them, donor participation is declining, and the pressure to do more with less has never been heavier. But the answer isn’t to hand your fundraising over to a bot.
In a recent webinar hosted by Avid, Kevin Peters (Founder & CTO, Avid) and Nathan Chappell (Chief AI Officer, Virtuous) made the case that the real opportunity in AI isn’t replacement — it’s reclamation. Reclaiming the hours your team spends on operational work so they can focus on the relational work that actually moves donors.
Why Most Fundraisers Have an Outdated View of AI
Most people’s mental model of AI is frozen at November 2022 — the month ChatGPT launched. Back then, it was a novelty: write me a poem, summarize this document. That version of AI feels familiar, and it’s largely how the sector is still using it.
But the technology has moved well past that.
Five months after launch, ChatGPT could pass the bar exam — outperforming 90% of lawyers. And according to recent research Nathan cited in the webinar, AI agents are on track to complete a full eight-hour workday without human intervention by March 2027.
What that means practically: the gap between how most fundraisers use AI and what AI is becoming is enormous — and it’s widening every month.
What Is an AI Agent — and Why Does It Matter for Fundraising?
An AI agent isn’t a single model. Kevin Peters described it this way: think of an agent as a manager orchestrating a team of specialized workers. When you ask it to do something, it delegates — pulling in the right tools, models, and data sources to complete the task faster and more accurately than any single model could alone.
This is how Edna, Avid’s fundraising agent, works. Rather than clicking through reports and building queries manually, Avid users ask Edna a question. She surfaces the answers, identifies the next best audience to reach, drafts segmented copy, and syncs lists to their email tool — in minutes, not days.
Nathan introduced a distinction worth internalizing: “human in the loop” (AI does the work, a human reviews) versus “human at the helm” (a human is ultimately accountable for every decision AI makes on their behalf). As agents become more autonomous, that distinction becomes more important, not less.
Watch a 6-minute tour of how Edna works as your team’s fundraising agent »
What Do Donors Actually Think About Nonprofits Using AI?
This is where the data is most surprising — and most useful.
FundraisingAI conducted a study of more than 2,000 donors and found:
- 80% of donors want nonprofits to use AI
- 90%+ say disclosure is key
But the more revealing finding was the disconnect between what donors want and what nonprofit professionals think donors want. Nonprofit pros assumed donors primarily wanted AI used for personalization. Donors ranked that fifth. What they actually want is for AI to help them trust the organization more — through better privacy, clearer accountability, and responsible stewardship of their data.
The practical takeaway: you don’t need an AI disclosure on every email. But you should have a clear statement somewhere on your website — similar to a data governance policy — that explains how you use AI and the principles guiding that use.
Is It Ethical to Use AI to Write Fundraising Copy?
Both Kevin and Nathan addressed this directly.
Kevin’s framework: treat AI output the way you’d treat copy from an agency. Would you send an email an agency wrote without reading it first? No. Would you let an agency rewrite your mission statement? No. Would you let them take an approved email and adapt it for an ad? Sure. The same logic applies to AI.
Nathan’s framing gets at something deeper: the ethical question isn’t who wrote the copy — it’s whether the donor is at the center of the decision. Fundraising has always involved persuasion. AI-generated copy isn’t inherently manipulative, and human-written copy isn’t inherently trustworthy. What matters is the intent and the values behind it.
His practical advice: write your organization’s AI values down. Most teams have values on the wall but don’t use them to make decisions. As AI introduces new ethical gray areas daily, internalized values become the framework.
How Should Nonprofits Protect Donor Data When Using AI?
Kevin’s rule is simple: if you wouldn’t want it repeated on the evening news, don’t put it in an AI tool.
More specifically:
- Free consumer tools (ChatGPT free, Claude free, Gemini) harvest your inputs. Never enter donor PII into these.
- Enterprise accounts (ChatGPT Enterprise, Claude for Enterprise, Microsoft Copilot with enterprise settings) offer walled-off environments where your data is not used to train models.
- When in doubt, anonymize: use donor IDs rather than names, hash identifying information before passing it to any model.
Enterprise-grade privacy that cost close to $1 million two years ago now runs around $30 per month.
At Avid, we go to great lengths to ensure no Donor PII ever reaches an AI. As Kevin shared in the discussion, there are creative ways to use tokens and IDs to anonymize your donor data.
Avid has this level of donor protection built in automatically. So you get all the benefits of better intelligence, analysis, strategy, and execution—and all the confidence that your donor’s info is secure.
Where Should Nonprofits Start with AI?
Both Kevin and Nathan pushed back on the instinct to start with the most visible or exciting use cases. The bigger wins come from auditing where time actually goes.
Nathan shared a framework one organization used: they catalogued 441 activities across their 150-person team, ran the list through Claude, identified which tasks AI could meaningfully augment, and then filtered for what no one actually enjoyed doing. They narrowed it to 20 priority areas — a concrete, actionable roadmap that came directly from their own work.
Kevin’s starting point for most fundraising teams: predictive AI and propensity modeling. One stat worth repeating — wealth is less than 10% correlated with generosity. Yet most organizations still use wealth as their primary prospecting signal. Machine learning can surface the engagement patterns that actually predict giving, and accessing that capability no longer requires a data science team.
Avid is built to be your AI fundraising agent out of the box.
Integrate all your tools with a few clicks. And Edna, Avid’s fundraising agent, is already trained on 8,000+ fundraising experiments and 650+ million donor interactions. She’s seen what works in fundraising across more touchpoints than any fundraising team sees in a lifetime.
Frequently Asked Questions
Can nonprofits use AI to write donor communications? Yes — with human review. Treat AI output as a first draft, not a final product. Always read, edit, and approve before sending.
Do nonprofits need to disclose when they use AI? Best practice is a clear AI use policy on your website. You don’t need to label every AI-assisted email, but transparency about how you use AI builds donor trust.
Is donor data safe in AI tools? It depends on the tool. Free consumer tools are not safe for donor data. Enterprise versions of major AI platforms offer strong privacy protections. When in doubt, don’t input PII.
Where should a small nonprofit start with AI? Start by auditing where your team’s time goes. The highest-value use of AI is eliminating the work that consumes hours and doesn’t require a human — not replacing the work that does.
Fundraising is human work. The teams that will win in this next era aren’t the ones that hand the most over to AI — they’re the ones that use AI to get back to the work only humans can do.