It’s that wonderful time of year again.  Somewhere, at this very moment,  amid the year-end scramble a fundraiser is opening the 2026 fundraising plan, changing the dates to 2027, adding 5% to the revenue goal, 7% to expenses, sprinkling some AI on top, and calling it strategic planning.

Don’t.  The ground underneath fundraising is moving too quickly for that.

Donors are changing how they find organizations. Artificial intelligence is changing how organizations find and communicate with donors. Fundraising platforms have become critical infrastructure. And while we have acquired astonishing new tools for acquiring donors, we remain remarkably talented at losing those donors.

So before the budget is finished, the consultants hired, the revenue targets engraved on tablets and the PowerPoint deck inflicted upon the staff, here’s a modest suggestion;put these four questions on the agenda.Consider it  The Agitator’s Checklist for Avoiding Yesterday’s Thinking.

☐ 1. WHAT HAPPENS IF THE MACHINERY YOU DEPEND ON ISN’T THERE TOMORROW?

Consider ActBlue. In the third quarter of 2026, ActBlue processed $833 million from more than 3 million donors making 19 million contributions. Its average gift was $44 and it was ActBlue’s best non-presidential quarter ever. Wonderful.

It is also a spectacular illustration of dependency.  At the same time all that money was pouring through ActBlue, the platform remained under intense political and legal pressure from the Republicans.  That doesn’t mean ActBlue is about to disappear.  It means fundraisers should understand a principle that every competent bank, hospital, airline and data center learned years ago:  Critical systems need redundancy.  Yet ask most nonprofits what happens if their primary donation processor stops working tomorrow morning and you may get something resembling the expression of a Labrador retriever encountering a screen door.

“We have another processor.”  Good.  Does it work?  Can you put up a donation page Monday morning?  Can you export the donor data?  Can your CRM talk to it?  Can you preserve recurring gifts?  Can finance reconcile the transactions?  Can your fundraising staff actually operate the thing?  There’s an enormous difference between having a backup and having a backup that works.  So put this on the 2027 planning agenda:  Identify every single point of failure in your fundraising operation.   Payment processor…CRM… Email platform…Data provider…Website…Analytics.   Then test the backup to  see  if any of these critical components fail. Not discuss it, or put it in a three-ring binder– Test it.

Success can create dependency just as easily as failure creates disaster.  ActBlue’s $833 million quarter doesn’t make contingency planning less important.  It makes it more important.

☐ 2. WHAT WILL AI SAY ABOUT YOU WHEN A DONOR ASKS WHERE TO GIVE?

For most of the digital era, organizations worried about Google and some spent fortunes on search-engine optimization.  We bought keywords with hopes of appearing on search engines’ page one.

But something fundamental is changing.  Increasingly, people don’t search,  they ask.  “What’s the best organization fighting Alzheimer’s?”,  “Which charity is effective at fighting hunger?”, “What organization should I support if I care about reproductive rights?”, “Which environmental groups can I trust?”   And an AI answers, maybe giving 5 organizations, maybe 3, maybe 1.  Is your organization in the answer?  More importantly what does that answer say  about you?  AI changes the communications problem in a fascinating way. For decades, fundraisers have obsessed over what we tell donors about ourselves, but now we also have to worry about what machines tell donors about us when we aren’t in the room.

And machines aren’t merely reading your beautifully polished case for support.  They may be encountering your financial information, independent evaluations,  news coverage, complaints, charity ratings, leadership controversies, government filings, our own website and whatever else the public record says about you.

Suddenly reputation isn’t merely a communications concern.  It’s fundraising infrastructure.  So sometime during your 2027 planning session, try something embarrassingly simple:  ask several major AI systems about your organization. Then ask them about your competitors and ask which organization they would recommend to a donor interested in your cause.

Ask which organization they would recommend to a donor interested in your cause and don’t forget to ask why.   You may learn more from those 15 minutes than from those expensive brand studies sitting on your bookshelf.

Because the next donor deciding whether to give you $100 may never begin that journey on your website.  She may begin by asking a machine whether you deserve it.

☐ 3. DECIDE WHERE AI STOPS BEFORE AI DECIDES FOR YOU.

AI adoption inside nonprofits is racing ahead of AI governance which shouldn’t surprise anyone.  Fundraisers have discovered that AI can draft appeals, analyze data, create donor segments, summarize research, personalize communications and perform in seconds tasks that once consumed hours.  Terrfic!

But there’s a question every fundraising operation needs to answer before somebody answers it accidentally:

Where is the line between personalization and impersonation?  Suppose AI drafts a first version of a donor letter.  OK.  Suppose it analyzes thousands of donor comments looking for patterns. Wonderful.   Suppose it suggests which donors may be interested in a particular program,, well that’s useful.

Now suppose it writes an intimate email to Mary about her late husband based on information buried in the CRM and sends it under the development director’s name—although the development director has never met Mary and never read the email.

Still comfortable?  Technology rarely announces the moment it crosses an ethical boundary.  It just makes crossing the boundary terribly convenient.

So I’d put one question in enormous letters on the wall during the 2027 planning meeting:

ARE WE USING AI TO HELP OUR PEOPLE HAVE MORE GENUINE RELATIONSHIPS WITH DONORS—OR TO MAKE DONORS BELIEVE A MACHINE IS A PERSON?

That’s not an anti-AI rule it’s a pro-AI rule. Use machines to remove drudgery, find patterns humans miss, improve drafts, help fundraisers spend more time being human.  But please don’t automate humanity and then congratulate yourself on personalization.  Remember: trust takes years to accumulate, and deception can vaporize it in an afternoon.

☐ 4. WHY ARE YOU SPENDING SO MUCH MONEY ACQUIRING DONORS YOU’RE APPARENTLY WILLING TO LOSE?

Here is one number I’d put on the screen during every 2027 fundraising planning meeting:

24%

That was the retention rate for newly acquired online donors in the latest M+R Benchmarks data.

Think about that. You acquire 100 new donors, celebrate while calculating the cost per acquisition and ask someone to make a colorful chart so your agency can present its stunning results.

Then 76 of those donors disappear and somehow that’s treated as a different department’s problem.

Perhaps we’ve been measuring the wrong victory because the same benchmark data contains another revealing number: Roughly 71% of new monthly donors were still giving after twelve months.

There’s a strategy hiding in the gap between 24% and 71%.  Yet look at most fundraising budgets and you’ll find that Acquisition has a budget, so does Media, Creative, Digital Advertising and Consultants.  But ow much money—and management attention—is specifically devoted to making a new donor want to give again?

And here’s the uncomfortable part: what do we reward? Gross revenue? Number of donors acquired? Cost per acquisition? ROAS? Campaign performance?  All perfectly reasonable measures.  But who gets rewarded because the donor is still there twelve months later?  If an agency acquires 100,000 donors, everyone applauds.  If 76,000 disappear, apparently that belongs to the future.

Maybe 2027 is the year to change that.  Maybe we should stop thinking exclusively about the cost of finding donorsand start paying much more attention to the value of keeping them.  Maybe retention, second-gift conversion and lifetime value should sit beside acquisition on every dashboard.

And perhaps some compensation should depend upon them; because a donor isn’t successfully acquired when the credit card clears.  A donor is successfully acquired when a relationship begins.

FOUR BOXES. CHECK THEM.

That’s it.  Before approving the 2027 fundraising plan, make sure somebody has seriously answered:

☐ What happens if our fundraising machinery fails?

☐ What does AI tell prospective donors about us?

☐ Where have we drawn the ethical boundaries around our own use of AI?

☐ What are we doing—and spending—to keep the donors we work so hard to acquire?

None of these will be the sexiest item at the planning retreat or staff meeting.  Nobody will leap onto the conference table shouting, “RETENTION POLICY!”  There probably won’t even be a standing ovation for redundant payment processing.  But all four questions are really asking the same thing:   Are we planning for the fundraising world that is arriving—or the one we already know how to operate in?

Because technology can find, score and predict a donor. It can also personalize a solicitation and process the gift and increasingly it can even tell the donor whether your organization deserves her money.  But there’s still one thing all this magnificent technology cannot yet manufacture: Trust.

Trust remains our job; so before you finish your 2027 fundraising plan you might want to make sure it’s actually in the budget.

— Roger