Dear Mary,

Thank you for your generous gift.

We know how much this cause means to you as we remember your grandson’s illness, your family’s experience, and the extraordinary loyalty you’ve shown over the years.

Your generosity means the world to us.

With warmest personal regards,

Your Development Director

Lovely note. Personal and heartwarming , but there’s one small problem.  The development director didn’t write it or read it.  In fact that devel0pement direct may not even know Mary exists.

The entire message was assembled by a machine that mined Mary’s donor record, identified the appropriate emotional triggers, and manufactured an expression of personal gratitude.  All in about three seconds.

And somewhere, a fundraising consultant, tech vendor or blogger is celebrating the enormous productivity gain.  Welcome to the brave new world of artificial intimacy.

THE DONORS HAVE SPOKEN. SOMEBODY TELL THE ROBOTS.

Last week, The Chronicle of Philanthropy reported findings from the third annual Donor Perceptions of AI study by Nathan Chappell and Cherian Koshy.

The researchers surveyed 1,020 American donors in September 2026.  Here are four numbers I’d suggest every fundraiser tape to the office wall:

• 70% are concerned about AI bots being presented as human representatives of charities.
• 77% say transparency about AI use is very important.
• 64% worry about losing the human connection with the organizations they support.
• 40% are uncomfortable with AI-personalized fundraising appeals.

And here’s the finding that ought to make every AI evangelist put down the PowerPoint clicker.  In 2024, 31.6% of surveyed donors described themselves as very familiar with AI.  By 2026, that number had climbed to 49.7%, yet the proportion uncomfortable with AI-personalized appeals remained virtually unchanged.

The donors learned more about AI. They didn’t necessarily learn to like it.

That rather spoils the comforting industry theory that donor resistance is merely a matter of education. But, perhaps donors understand perfectly well what the technology can do and perhaps that’s precisely what worries them.

INTRODUCING YOUR NEW DEVELOPMENT OFFICER. PLEASE DON’T ASK TO MEET HER.

While donors were expressing these reservations, the fundraising technology industry was busy unveiling the next generation of automated relationships.

Version2.ai, associated with Givzey, markets a Virtual Stewardship Officer that promises personalized donor journeys conducted autonomously.  The system is designed to provide ongoing stewardship, maintain donor engagement, and encourage renewals.

The company’s ambition is to make meaningful stewardship available to donors regardless of gift size and this is certainly a worthwhile ambition.  Most organizations do a miserable job of staying in touch with smaller donors except when they want another contribution.

If technology can help correct that neglect, terrific.  But the phrase autonomous stewardship deserves examination.   Stewardship has traditionally meant recognizing a donor’s generosity, demonstrating the difference her gift made, and building a relationship grounded in mutual commitment.

Now we’re proposing to automate the relationship itself. Which raises a question:  If a donor believes she’s developing a relationship with someone who doesn’t actually exist, is that stewardship—or theater?

Meanwhile, and also last week, Bonterra introduced Que for OneCause, an AI coaching system drawing on information from more than 96,000 fundraising events and fifteen years of event experience.  Que advises fundraisers on auctions, event strategy, donor engagement, and follow-up.  Importantly, Bonterra describes this as assistance for human fundraisers, not a replacement for their judgment.

And that distinction matters enormously.  One system helps the fundraiser decide what to do, the other offers to carry out ongoing donor stewardship autonomously.  Both may have legitimate uses.

But they represent different answers to a fundamental question:   Are we using AI to help people build relationships—or to replace the people who build them?

PERSONALIZATION ISN’T THE SAME AS BEING PERSONAL.

Let’s return to Mary and the thank you note.

Mary has supported a children’s hospital for fifteen years because the hospital helped her grandson survive a serious illness.

Her giving history is in the database, so are notes about her family, her interests, and perhaps the anniversary of her grandson’s treatment.  And so an AI system can assemble all that information and produce a lovely message like mentioning the grandson, expressing warm appreciation for Mary’s loyalty and, of course, even sign the development director’s name.  And it can do all this in approximately the time it takes a human being to locate the stationery.

NOW… imagine Mary discovers that the development director never wrote the message, never read it, and perhaps never even knew it was sent.  Would Mary feel appreciated…or processed?  Frankly, I don’t know. Perhaps Version2.i.ai does.

That’s why the new donor research matters.  It doesn’t establish that every AI-assisted communication damages trust, nor prove that donors will stop giving when they encounter automation.  But it tells us something we all should take seriously: Donors care whether they’re dealing with a person or a machine.  And many don’t appreciate being left to figure out the difference.

The issue isn’t simply whether the information in the message is accurate; it’s whether the apparent relationship is authentic.  That the computer can remember Mary’s grandson doesn’t mean it remembers Mary..

DON’T CONFUSE THE APPEARANCE OF ATTENTION WITH ATTENTION ITSELF.

The most potentially dangerous word in fundraising may be scale.  Scale acquisition… scale personalization…scale stewardship…scale relationships.

Why spend twenty minutes writing a thoughtful donor note when software can produce a thousand or more in the same amount of time?  The answer, of course, depends on what we’re trying to accomplish.  If we’re trying to deliver accurate information, remind donors of important developments, acknowledge transactions promptly, or make sure nobody falls through the cracks, automation can be a godsend.

But if we’re trying to convince 20,000 people that someone personally cares about each of them when nobody actually does, we have a different problem.  We’ve confused the appearance of attention with attention itself.   And that’s a dangerous confusion in a trade built on trust.

The temptation is understandable because development departments are understaffed, some donor files are enormous, retention is chronic problem that brings with it the relentless pressure to raise more money from fewer people.

So when a vendor promises personalized stewardship at scale, the prospect is irresistible.  But before signing the contract, perhaps somebody should ask:  Does this technology create a better donor experience—or simply a cheaper imitation of one?

Those are not necessarily the same thing.

EVEN THE DONORS WHO LIKE AI WANT TO KNOW WHO’S BEHIND THE CURTAIN.

The Chappell-Koshy study contains another finding worth considering.  Among donors who said they were very familiar with AI, 85% still considered transparency very important.  I find that fascinating because the survey respondents weren’t necessarily people frightened by a technology they don’t understand.  In fact, they may be perfectly comfortable using AI themselves.  They simply want to know when a charity is using it on them.

And there’s a lesson here for fundraisers who believe disclosure will somehow ruin the magic.  If telling donors how you’re communicating with them ruins the relationship, perhaps the relationship wasn’t quite what you thought it was.

Bennett Weiner, CEO of the BBB Wise Giving Alliance, [Full disclosure: of which I’m a board member] makes a sensible point in The Chronicle: organizations should be open about AI so donors can make informed judgments rather than discover its use as an unwelcome surprise.

Imagine applying that principle to your work:

• Tell donors when they’re interacting with an AI assistant.
• Explain what information about the donor the system uses.
• Make it easy to reach a real person.
• Don’t send machine-generated expressions of intimate personal concern under someone’s signature without meaningful human involvement.
• And don’t confuse an AI disclosure buried in the privacy policy with genuine transparency.

None of this requires abandoning AI.    It requires abandoning deception.

THE BEST USE OF AI MAY BE TO GIVE FUNDRAISERS MORE TIME TO BE HUMAN.

I use AI, enthusiastically because it helps with research, analysis, drafting, testing, editing, and the thousand small chores that consume time better spent thinking.  This technology is an instrument, not (yet) a substitute for judgment, and certainly not a substitute for caring.

Here are some suggested uses you might try that will save you time and boost your bottom line:

• Imagine a development department that uses AI to identify donors who haven’t received appropriate acknowledgment.
• To flag supporters whose questions have gone unanswered.
• To prepare useful background for a genuine (human-to-human conversation.
• To summarize program accomplishments so you or another fundraiser can explain them clearly.
• To discover that Mary hasn’t heard from your organization since her last gift.

Then imagine an actual person picks up the telephone and calls Mary, listens, asks about her grandson, and means it. That’s AI making fundraising more human.  Not because the machine has learned to care, but because it’s  given someone who does care the time and information to demonstrate it.

THE THANK-YOU NOTE THAT NOBODY WROTE.

I suspect we’re approaching an uncomfortable moment in fundraising.  For decades, we’ve told ourselves that relationships matter, that donors are partners, not ATMs, that appreciate must be heartfelt, and timely.  In short, loyalty is earned.

Now technology offers to automate much of what we’ve spent years describing as fundamentally human and the industry is understandably excited.  But the new donor research suggests we should be careful about what we automate—and especially about what we pretend not to automate.

Because there’s a difference between a machine helping someone express gratitude and a machine manufacturing gratitude on someone’s behalf.  The first can strengthen a relationship.  The second may eventually destroy it.

Here’s the irony—and it hits me each time I read one of Kevin’s posts on experiments and practices over at DonorVoice using all kinds of data and techs but with a finely honed sense of the truly important:  we have never had more powerful tools for learning about donors, understanding their interests, and communicating with them.  Yet we risk using those tools to create relationships in which no human being is actually present.

Perhaps the most important question for the next generation of fundraising technology isn’t how convincingly a machine can imitate a person.  It’s how much more time the machine can give a person to do the things only a person can do.  Donors don’t need us to manufacture intimacy.  They need us to earn trust.

Gratitude isn’t something you generate. It’s something you feel. And donors deserve to feel the difference.

Roger