AI is a new kind of fire. How will human fund development professionals respond?

Let me start with a confession. I am deeply skeptical of AI in fundraising.

As a novelist and composer, I try to steer far clear of it. It wants to destroy my human creativity, and I won’t let it. I compose at a piano, using pencil and paper. I outlined my novel on a whiteboard, using an erasable marker, and I write it in Obsidian, an app that doesn’t force AI on me like Word does.

In our field of fundraising, steering clear of AI is not so simple. The major CRMs used by organizations to keep track of their donors have been adding AI, whether we want it or not. It’s creeped into countless other areas that impact nonprofits, too.

Fundraising professionals need to become, if not experts, at least professionally competent in the ethical use of AI in fundraising, because, used improperly, it threatens to damage the human relationships that make everything else possible. If we use AI the way the hype cycle suggests, as a shortcut for producing more appeals, more emails, more “personalized” communications at higher volume, we will increase the distance between ourselves and our donors. We will produce more words and build fewer relationships. Our donors will feel it. They will sense that something mechanical has replaced something human, and they will drift away.

I wrote my book, Holding Fire, because of my deep skepticism, but also because, properly understood and with the right approach, it can help us humans devote more time to building the very human relationships so vital to a successful fund development program.

Why me? I’ve spent decades studying AI theory and developments while also building, administering, and advising fund development programs, not to mention publishing dozens of books on the subject through CharityChannel Press, the publishing house I founded. I saw this moment coming, years before ChatGPT brought AI to the masses.

Introducing the Human-Centered AI Framework

Somewhere in our distant past, a human being picked up a burning branch. Those early humans discovered fire could warm them through a cold night, cook food, light the darkness, keep predators at bay. They also discovered it could burn their shelter to the ground, scar their children, and destroy everything they’d built. Fire was the most formidable tool early humans ever encountered, and it demanded respect, attention, and constant vigilance. The communities that thrived were the ones that learned to hold fire carefully, to use it with purpose, and to never forget what it could do if they got careless.

We have just invented a new kind of fire. The difference between AI that strengthens donor relationships and AI that corrodes them comes down to how we use it. To make sure we use it the right way, the ethical way, requires a framework that keeps the human at the center of every decision, every communication, and every relationship. It treats AI as a tool that serves our ability to connect with donors, the way fire served early humans’ ability to survive and flourish, as long as they held it with care.

In preparing to write my book, I developed the Human-Centered AI Framework. It has five elements. I want to walk you through all five here, because once you see them, you’ll recognize them in every piece of fundraising advice that follows in this series. They also govern everything in the book. My goal is that you’ll have them govern everything you do when you employ AI in your fund development work.

What AI Does Well, and What Remains Beautifully Human

If we give AI a donor’s giving history, conversation notes, event attendance, and engagement patterns, it can identify trends, spot opportunities, and draft a personalized communication in seconds. It has the patience to read every note, cross-reference every data point, and never skip a donor because it’s running late for a meeting. These patterns exist in our data right now. We just can’t easily see them because no human being has time to analyze hundreds of relationships simultaneously.

The relationship itself, though, belongs entirely to us. AI can prepare us well for a conversation, and the preparation it provides is remarkable. The conversation itself, the listening, the judgment, the emotional intelligence, that’s ours. When we sit across the table from a pediatrician donor and she tells us about the children in her practice who come in hungry and what it does to her as a physician, that moment of human connection is irreplaceable. AI got us to that table well-prepared. What happens at the table is entirely, beautifully, human.

The first element of the framework is this healthy division of labor: AI handles synthesis, drafting, pattern recognition, tracking, and portfolio-wide analysis. We humans handle relationships, judgment, emotional intelligence, program knowledge, and every strategic decision about when to push forward and when to pull back. AI amplifies our capacity to do the human work.

The second element is the workflow that makes the division of labor practical. Every interaction in the system follows the same cycle. The AI scans the donor base and reveals what needs attention, prioritized and with a specific rationale. It drafts a communication or briefing. We review the draft, edit it, add the personal knowledge and program details that only we possess, and make the strategic call about what to send and when. We send the note, make the phone call, schedule the visit. Then we update the donor’s record with what we learned, enriching what the AI will know for next time.

The cycle becomes second nature quickly. Imagine the AI notices that a donor’s giving has increased three times this year, each gift larger than the last. It drafts a personal thank-you that acknowledges the trajectory. You review it, trim the language, add a coffee invitation because you’ve been thinking about showing her the mobile pantry in action, and send it. Four minutes. The AI gave you the draft. You made the relationship move that only a human would think to make.

A System That Learns From Us

The third element, the virtuous data cycle, is the reason this approach is fundamentally different from a one-time AI tool you use to generate a letter.

When you edited that thank-you and added the coffee invitation, you made a judgment call based on your personal knowledge of the relationship. If you take 30 seconds to note that decision in the donor’s record, the AI knows about it next time. The next draft it produces for that donor will reference the invitation. The draft after that will build on whatever happened at the coffee meeting, because you’ll have logged that conversation, too.

We add context the AI could never generate on its own: program details from last week’s site visit, a personal observation from a gala conversation, a strategic judgment about timing. That context enriches the donor record. The richer record produces better AI output. Better output means less editing time and more time for the relationship work only we can do. The system improves with every use, and the improvement compounds over months and years. A letter generator gives you the same quality output on day one and day three hundred. This system gets more personal, more accurate, and more strategically useful the longer you use it.

The fourth element is the scale shift. For decades, the kind of thoughtful, well-prepared, personally informed fundraising I’ve been describing has been reserved for a handful of top prospects. A major gift officer working with 20 donors could do deep preparation for each visit, maintain detailed relationship notes, and provide individualized stewardship. Everyone else got the form letter.

The framework changes that. When AI handles the scanning, drafting, and tracking, our capacity to provide individualized attention extends across the entire donor base. A brand-new $50 donor gets a welcome sequence tailored to specific interests and family connections. A faithful mid-level donor who has given every year for six years gets a phone call informed by everything the system knows about a personal connection to the mission. A major gift prospect gets a pre-visit briefing that synthesizes months of relationship intelligence into the best-prepared meeting of her philanthropic life. The same quality of preparation and attention can now reach every donor at every level, because AI handles the production work so we can do the human-driven fundraising.

If you see silos in the development office dissolving, you’re not alone. I do, too.

Holding Fire Carefully

I saved the fifth element for last because it governs all the others: permanent ethical vigilance.

Remember fire? AI is fire, and we are playing with it.

AI can write a convincing personal thank-you note. It can also craft a solicitation that pushes exactly the right emotional buttons at exactly the right moment. It is here that we need to be especially careful, because the same capacity that helps us build a deeper relationship could be used to manipulate a donor into a gift the donor wouldn’t otherwise make. The line between personalization and manipulation can be subtle.

In the framework, every communication the system produces, every briefing it prepares, every pattern it shows us, should pass through a simple test: Am I using what I know about this donor to advance this donor’s philanthropic goals, or am I using it to advance my revenue targets? When those two things align, as they usually do, there’s no tension. When they diverge, the donor’s interest must come first. Always.

It would be foolish to adopt AI in fundraising without maintaining a deep and permanent skepticism about how it’s being used. AI doesn’t know where the ethical lines are, not really. It’s up to us humans to be fully present when we’re working with it, and to keep it in check when it oversteps itself.

These five elements, a healthy division of labor, a repeating workflow, a virtuous data cycle, a shift in scale, and permanent ethical vigilance, are the Human-Centered AI Framework. They are the foundation of everything in my book. Every chapter is an application of this framework to a specific fundraising challenge: annual campaigns, major gifts, stewardship, lapsed donor recovery, board engagement, and the ethical questions that tie it all together.

The framework keeps the human at the center of every decision, every communication, and every relationship. AI empowers us to do more of what only we can do, with more donors, more consistently, and with better preparation than was ever possible before, as long as we hold it with care.

Next week, I’ll show you the framework in action during a 20-minute Monday morning review that ensures no donor in your portfolio slips through the cracks.