AI won't fix fundraising. It will simply feed on what's been broken for years

How AI is going to impact fundraising. Why it is likely to take your job and why most of the sector is having the wrong conversation about what it all means.

This is going to be about AI (obvs). But it is not a long version of those LinkedIn posts written by agencies that are basically saying you’d best work with them on using AI or you are going to miss out.

The fundraising sector has always had a serious case of FOMO, and that is something agencies are apt to exploit. Someone presents a new idea at a conference and, even though it is still early days, huge numbers will be mentioned that offer the promise of a step-change in income. Who remembers all the fun and games around engagement products or text giving? Or maybe a campaign wins an award. Or a press report on a rebrand announces that a new logo and colour palette have lifted propensity to give and brought in loads more supporters.

Suddenly everyone with a budget wants to know more. Partly so they don’t look like they are behind the curve, and partly because they’re hoping it might finally point them towards that mythical magic money tree dripping with cash. Then within a year or two, when hints of the truth comes out, the new idea turns out to have been a disaster that acted as a distraction, or the rebrand was found to have caused a downturn in income because people who previously knew and loved the charity became confused and felt pushed away.

And now we have AI. And it has a sniff of that sort of project about it. But only a sniff. That’s because AI is different. It isn’t just a repackage of something we’ve always done, dressed up in a slightly different way. AI could fundamentally reshape how charities operate in society. Notice that I didn’t say fundraising there. That’s because what we might be looking at ten or twenty years down the line may not actually be fundraising in the sense we know it today.

And those friendly LinkedIn posts, white papers, downloadable decks and conference sessions that present AI as a way to be more creative or to make fundraising a bit more fun or to be all lovely and innovative all miss the point. I don’t think anyone outside a handful of fundraisers (big shout out to Denisa Casement) has really started to understand how much it will change what we do. AI is not a little toy. It is something that could make the sector unrecognisable in a relatively short period of time.

Let’s consider the framing

The dominant framing seems to be one of productivity. AI is presented as a tool that helps fundraisers do more, faster, cheaper. It will write your appeals for you, it will analyse your results and tell you what that means. The secondary framing is moral. AI must be adopted responsibly, with the right guardrails. That seems to mean that we must keep humans in the loop, protect beneficiaries, build a code of conduct and run a values workshop. Both frames, in my view, miss what is actually happening.

I’ve just returned from giving the opening plenary at Congreso Fundraising in Spain, where I shared my thoughts on the enshittification of fundraising, a phenomenon that continues to gather pace. I’m presenting on this in the Netherlands in September, a small number of large INGOs have asked for private sessions on the same theme, and it will be a big part of my Australia and New Zealand tour at the end of July.

My fear is that AI could further turbo-charge the enshittification of the sector. That is something we should choose to guard against. The fact is, AI will expose and exploit the fundraising sector’s existing and ongoing weaknesses – its preference for shiny balls, innovation and the promise of rapid growth over patient relationship-building.

That is what lies at the heart of the problem. AI gives us the ability to do the very things that have already eroded trust, only faster, cheaper, and at greater volume. The result will be more asks. More automation. More distance between the donor and the people the charity is working with. Which brings us to you, dear fundraising friend. What impact might AI have on you? Let’s have a look at the evidence.

The productivity story doesn’t seem to survive contact with the product

As I write, the AI conversation is playing out in real time at the CIOF Fundraising Convention. The Friday opening plenary was given by Blackbaud, talking about AI, Trust, and the Future of Supporter Experience: The New Era of UK Fundraising with new UK research and what the convention materials describe as, “an honest conversation about what it means for our profession”. Across the programme there is a small selection of other takes on AI. It is almost as if AI is a sub-topic, not a category. A way to make our jobs a bit more fun, a bit more rewarding and a bit more effective.

But then look at what Blackbaud’s CEO, Mike Gianoni, was reported in MarketBeat as telling investors at JPMorgan a short while ago, and a slightly different story emerges.

Blackbaud has launched its first “fully agentic” product, a fundraising agent that is reported to have been generally available for six weeks at the time he sat down with JPMorgan’s technology banking team. They have called the category, with no apparent irony, Agents for Good. The product appears to use Blackbaud’s data, predictive analytics and wealth screening tools to profile potential donors, generate comms via SMS, email or an avatar, and then either hand a larger opportunity off to a human fundraiser or close the donation transaction itself.

It is reported as being priced at a flat $25,000 to $30,000 a year on a fixed annual fee, with the ROI tied to how much the agent raises. The example shared was an American university with 190,000 alumni whose staff can only ever reach a fraction of that audience. An AI fundraising agent could scale outreach to donors who would otherwise not be contacted, including recent graduates who might start as small monthly donors. It was reported that Blackbaud is running webinars with hundreds of existing customers and signing new ones every week, across higher education, hospitals, schools and nonprofits.

As Blackbaud themselves describe it on their website:

“That’s the promise of agentic AI–a new kind of intelligence that can plan, decide, and act on its own. For social impact organizations like yours, this shift is exciting and transformative.

Agentic AI has the potential to automate complex workflows, personalize engagement at scale, and give your team more time to focus on the human connections that drive your mission forward.”

So one of the most widely used fundraising platforms in the sector is actively selling what looks very much like an autonomous AI fundraising agent. It appears to be able to profile donors, build communications, shape the ask, contact supporters automatically, and, in some cases, help move the donation process through to completion, all under the reassuring phrase “human oversight”. And that company is giving an opening plenary at the UK’s flagship fundraising conference on “trust” and “the future of supporter experience.”

The conversation the sector is starting to have, with its codes of conduct and its humans in the loop, is happening at one level. The product being sold into the sector is operating at quite another. For the equivalent cost of developing the creative for a single mailing pack with a decent agency, you can now buy access to an autonomous agent that identifies the donors, communicates with them, and completes the donations on its provider’s payment rails. The exploratory conversation is busy listing principles. The product is in market, signing customers weekly, at round about $25-30k a year. And, if I am honest, if you are not very good at fundraising it’s probably going to give your income a boost.

But. And it’s a big but. Are we are supposed to believe that fundraisers will use the time freed up for strategy and human connection? We will, as they say, have to wait and see.

The fundraisers who worry about AI seem to be reading it right

Let’s dig a little more deeply into some of the main issues being raised about AI. A number of fundraisers I know, particularly the experienced ones who have spent decades thinking about how their work actually affects donors, have been questioning the adoption of AI for some time. Like me, many value some of what it offers. I personally love it for research, for creating copyright-free images, and for building charts from raw data. All things I am doing for myself, in my own work, where the output is going to be checked and judged by me. But I don’t think any of them fully trust it. The more I talk to them, the clearer it gets. They think something significant is happening that the sector hasn’t fully considered. I think they are more right than wrong, and “don’t worry, we’ll have a code of conduct” is not the answer we need.

Much of the debate is focused on the fact that AI isn’t very good at being creative. To my mind that’s true. It really can’t come up with a decent fundraising concept and its copy rarely rises above OK. But then again, I know it will get better. Compare what AI is delivering now with a year ago, and there has been a step change. This will only improve.

But we also have the energy and water question, which is one that gets cited regularly. The figures we often see are mostly out of date, and per-query energy use has actually come down over the past two years. But the direction of travel is clear. The new generation of reasoning models use significantly more energy than the tools the early stats described. ChatGPT currently serves around 2.5 billion queries a day. The IEA projects global data centre electricity demand to roughly double by 2030. And cooling water draws are concentrated in regions that can least afford them, places like Spain and Arizona. Any individual charity’s footprint is tiny. But as a sector, it is going to add up.

There was a long piece in the Financial Times this week about Anthropic, the company that makes Claude. It gave a sense of what the morality conversation looks like among the people building this technology. Anthropic was founded in 2021 by researchers who quit OpenAI because they were worried about how fast AI was being deployed. They built a deliberately unusual structure. The company itself is a public benefit corporation, legally required to balance profit with social impact. Its directors are appointed by a separate body, the Long-Term Benefit Trust, whose trustees have no financial interest in the business. It has refused US Department of Defense contracts at significant commercial cost. All seven co-founders have pledged 80% of their wealth to fund social good. But, and it is a big but, even then its CEO, Dario Amodei, has said publicly that he thinks there is a 25% chance AI leads to catastrophe for the human race. And even Anthropic, careful as they have been, are still being criticised by serious people in the field. Connor Leahy, who runs an AI policy non-profit, told the FT: “If you still find yourself building this technology, then you’re the bad guy.”

This is the real ethics conversation about AI. It is not “should we have a sector code of conduct.” Our focus should be on what we’re actually using, the trade-offs behind it, the long-term consequences, and what responsible use means when even the developers admit they don’t fully understand how their models will behave. Anthropic appear to have built something they think is a 25% bet on civilisation – and they really do seem to be putting safety first. The companies racing them have not built anywhere near the same governance scaffolding. The official conversation in fundraising, by contrast, consists mostly of “let’s keep a human in the loop and write a code of conduct.” It is operating at a level the actual stakes don’t begin to allow.

What this means in practice is that the experienced fundraisers who instinctively dislike AI are picking up on something that isn’t featuring in enough discussions. It isn’t that AI is going to necessarily become the Skynet of Terminator fame. It is that there is a real, moral question about what you are plugging into the heart of donor relationships when you start letting these tools write the appeals, generate the storytelling, simulate the conversations. I don’t think the developers themselves really know the implications of what they are building. The sector’s response is to recommend you use it cautiously while running a workshop on values. That is not adequate to the actual situation, and the fundraisers who feel that in their gut are probably reading it more accurately than the people writing the codes of conduct.

It also means the real question for a fundraiser isn’t “are we using AI ethically?” It’s harder than that, and it comes in several parts.

Do you actually understand what you are putting at the centre of your donor relationships? Do you know which AI providers are taking the safety question seriously, and which are racing for revenue with the brakes off? Are you comfortable with what this means at scale - not just for your charity, but for the donor on the other end, who will increasingly be unable to tell whether anything they receive was made by a human at all?

And do you have a view on what happens next? On what becomes of all that data for example, who supports which charity, which progressive cause, at what level, and when? On who might, one day, buy Blackbaud, and everything that comes with it?

For most charities, the honest answer is “no, and we haven’t really thought about it.” Which is exactly the conversation we should be having.

The central dynamic: production becomes free, attention does not

OK. Having covered the simple stuff, how AI might destroy the world and fundraising as we know it, let’s look at the impact on donors.

Direct response fundraising has historically depended on the fact that learning took time. Writing a great appeal needed skill. Finding the right people to send it to was hard and expensive. Testing what worked for your audience was a significant task. The charities and agencies that did those three things well had a defensible edge, because many charities, as I regularly point out to my regular readers, did them badly.

But here we are and all three are now available pretty much for free. Anyone can produce an OK appeal in under an hour. Anyone can build a lookalike audience. Anyone can run ten A/B tests in the time it used to take to run one. On the face of it, that sounds like AI democratising fundraising. Everybody wins. But it raises a more interesting question, which is what happens when everyone does this at the same time? The risk is a tragedy of the commons. And the commons in question is donor attention.

When the cost of producing a “good enough” email appeal collapses to zero, every charity can easily produce more of them. The supply of asks goes vertical. The number of inboxes and feeds available to receive them does not. A mid-sized charity that used to send half a dozen email appeals a year, because each one required a copywriter, a designer, sign-off and a project manager, now sends six a month. But it isn’t alone. Every other charity can do the same. The donor who used to receive twenty emails a year from organisations she is vaguely connected to is going to receive two hundred and fifty. She is going to do what every human does in that situation, which is to stop opening any of them.

Cheaper appeals do not necessarily produce more donations. They produce channel saturation, which may speed up the death of email and social media as viable mechanics. The first charities to lean hard into AI-generated mass appeals will get a brief sugar high, and probably win a few awards. The ones who follow get nothing. The donor stops reading. Hold that in mind for everything that follows.

Send better emails, not more

Speak to donors about email and a clear age split emerges. Younger donors are actively telling us they ignore it. They sign up because they have to, often to get an activation code for an app or a subscription. Older donors are still scrolling through their inboxes. They still use email, and they still read it. For older donors, fundraising emails still work, for now.

To keep them working, charities need to consider two things. Deliverability, and attention. Both seem to be eroding fast.

Deliverability is getting harder because the major inbox providers are tightening up on bulk senders, low-engagement mail and anything that looks repetitive, synthetic or unwanted. AI makes it much easier to produce exactly that kind of volume – endless variants of the same appeal, sent more often, to people whose attention is already stretched. Gmail does not need to know that a message was written by AI to send it to Junk. It only needs to see the familiar signals of bulk fundraising mail that people do not open, do not click, delete quickly, or mark as spam. The same technology that helps charities generate more appeals also helps inbox providers filter, classify and bury them. And in that contest, the inbox provider has the home advantage.

Attention is dying for the reasons above. Even the appeals that land in the primary inbox are competing with many others. Open rates are shrinking. Much of the sector’s response, predictably, is to send more email. This is approximately the response of a man digging a hole, being told to stop or he won’t be able to climb out and concluding he needs a bigger spade.

The answer to this is to do the opposite. The charities that will keep email working are the ones that send better emails, not more, and that build something more like a personal email with stuff added to it – created with real DM skill. Or a newsletter people actually want rather than a sequence of asks. This is what every sensible, digitally aware fundraising consultant has been saying for ten years or more. The difference now is that the penalty for not listening is going to be much more expensive.

Paid social and search

Paid acquisition through Meta and the rest has been a reasonably effective channel over recent years. Intelligent charities with good creative approaches could acquire donors at a reasonable price and start returning a profit within a year or so.

But when OK creative is free, that advantage is available to everyone. Every charity has reasonable creative. Every charity is bidding on the same audiences. Meta charges more, and makes more. The cost per acquired donor goes up. The lifetime value of the acquired donor goes down, because she is more expensive to acquire in the first place, and because she is now being asked by everyone else. The maths stops working. Many organisations whose growth strategy has been “scale paid acquisition” are going to discover that the strategy stopped working sometime in the last two years, and they hadn’t noticed, because the dashboards were built on a model of the world that is two years old.

There is also a second problem. The platforms themselves are using AI to optimise ad delivery, which means they are optimising their profit. If Meta’s algorithm can predict exactly how much your charity is willing to pay for a donor, and charge you that amount, your margin could well be on it’s way to zero. This is already happening in commercial e-commerce. And from what I hear, it is starting to happen in fundraising too.

Social media has an authenticity problem

Organic social has been a problem for years, and it doesn’t look like it’s going to get any better. AI accelerates this issue in two ways. Feeds get flooded with AI-generated content, which forces the platforms to push organic reach down further to make room for ads. And, worse, everything becomes suspected of being AI-generated, including the things that aren’t.

This is the bit the productivity framing won’t touch, and it is also where the morality question starts to bite. A charity posts a photo of a real beneficiary telling a real story, and a growing percentage of viewers now wonder whether the person is real, whether the story is real, whether the whole thing was generated by an agency in a different country.

The default assumption shifts from “this is true unless I have reason to doubt it” to “this is fabricated unless I have reason to believe it.” Trust runs on the assumption, not on the evidence.

The charities that get ahead of this will lean hard into provenance. Live video. Unedited footage. Long-form content where the same beneficiary or staff member appears repeatedly over years. Local fundraisers where the supporters know the staff by name. Anything that is hard to synthesise and easy to verify. The polished ninety-second video with the swelling strings and the perfect lighting becomes the least trusted format in the mix. The exact opposite of what the sector has been optimising for since 2005, and the exact opposite of what those session on “AI for sharper communications and storytelling” tells you to do.

The money is old

So far this has been a story about digital channels. Here is the point where we move beyond that to something I have been banging on about for years, because for a long time nobody else seemed to be. Older people are the past, the present and the future of fundraising. I am pleased to say that some agency types are finally starting to wake up to this, after years of selling rebrands and new products aimed at Millennials. I am seeing more agency names in my subscription lists, and more agency connection requests on LinkedIn. Particularly from the ones who have spent years telling the sector that Millennials are where YOU need to spend YOUR money with THEM.

The donor base for UK individual giving skews heavily toward people over sixty. The same is true in the US, Canada, Australia, New Zealand and Western Europe. This is not a small bias. It is the entire shape of the sector. A charity getting £20m in individual giving is almost certainly getting most of it from people whose primary news source is a printed newspaper and the evening news, and whose feelings about TikTok range from puzzled to suspicious.

Which means the channels that actually move money in the sector are direct mail, direct response television, radio, print press, telephone, and word-of-mouth through churches, clubs, and community groups. The hand-wringing about Meta costs and email open rates is real, but it is hand-wringing about a channel mix that was producing low-LTV younger donors anyway. Losing it hurts if that’s where you’ve bet the farm, but it is not where the money is.

So what is happening to the older-donor channels? They are being disrupted too, but on a different timeline and through different mechanisms. None of them are getting saturated the way social and email have, because the cost structure of physical mail and broadcast media puts a natural ceiling on volume. And although this group is getting older, they have plenty of life left in them, and their wealth is growing.

The Boomer generation and the older Gen Xers, sometimes called Generation Jones, are the last cohort that give as a default civic act. Younger donors (outside people of faith) give less reliably, and with less institutional loyalty. This is particularly true of Millennials, who give differently. To specific causes, often via crowdfunding, often as one-off transactions rather than sustained relationships. The pipeline that was supposed to refill the donor pool continuously is starting to look like it is running a bit dry. But that is the topic for a different post.

DRTV and direct mail

Good direct response television has been a reliable workhorse for UK charity acquisition for thirty years. It still works, when it is done well. But TV advertising has not been immune to the enshittification bug. I have written elsewhere about what I call the John-Lewisification of charity advertising, where increasingly the ads are built around a clever – if obtuse – idea rather than human connection, for cash and RG recruitment and legacies. When I sit down with donors and we watch these ads in research, donors tell us, plainly, that they do not work. An ad that copies John Lewis is not an ad that raises much money.

Good ads, though, still recruit donors. Particularly when they are relevant, and particularly when they ask for cash rather than a £3-a-month direct debit. But they are working from a shrinking base. The people watching linear broadcast television is also shrinking.

The streaming services that have captured the displaced eyeballs do not deliver the same DRTV economics. You can buy advertising on various ITV services and Talking Pictures TV, but the captive-audience dynamic of a 75-year-old half-watching Countdown in the afternoon and seeing the same charity ad eight times in a fortnight is not reproducible on Netflix.

Direct mail is the far more interesting and exciting. That’s because it actually increases in value in an AI-saturated world, not less. The reason is that the channel is bandwidth-constrained in a way that digital channels are not. Royal Mail will not be delivering four hundred letters a day to your house. The cost per piece is still meaningful enough that you cannot blast infinite variants. And, this is the part that few saw coming five years ago, a physical letter from a real person in 2027 will feel like an artefact in a way it didn’t in 2015. When everything digital is suspected of being AI-generated, correctly, something that arrived by post, with a stamp, with a real signature, that doesn’t look like it was created for a bank, carries a weight it was presumed to have lost decades ago.

The charities that abandoned direct mail in the 2010s because email was cheaper are going to look at their lapsed donor data in 2028 and feel more than a bit foolish. Because for retention, reactivation, and stewardship of donors over fifty-five, the physical letter becomes the most trusted touchpoint a charity has. Since I set up MPFA, two of my biggest jobs have been working with charities to revitalise their DM programmes from the ground up. And just this week I had my first conversation with a senior Spanish fundraiser that started with “we want to get back into mail.” They will not be the last.

What gets valuable

Three things that will help pull you through the mire of AI content…

Connection. The transactional appeal, “give now to help X,” is what gets crushed by saturation. What survives is connection and engagement. People who care about an issue and feel their giving gives them a sense of agency to do something good. (Oh, the irony.) They feel they belong to something. They are known. They get an annual report that isn’t pretending to be one. It is not so much about an ask as an extension of their identity.

Thanking. Charities that treat their existing supporters like a precious resource will keep them on board. This means thanking properly, and taking it offline. Again, anything that feels real. In this world, authenticity is going to hammer commercial-style brands. Thanking means you need to flip the standard fundraising economics that has characterised the sector for years. It’s not longer good enough to recruit and hope. Now it’s about finding the right type of donor and working bloody hard to keep them. It’s think where that relationship could go. And with older donors, that is usually legacies.

Specific, local, visible work. It is probably the mass fundraising at the national level that is going to struggle, unless strength is part of your brand, à la UNHCR or World Food Programme or Cancer Research. Hyper-local fundraising such as the food bank, the hospice or the kid’s sports club gets stronger, because the donor can verify the work with her own eyes and doesn’t need the appeal to convince her of anything. The decade ahead may not be kind to the £200m international generalist NGO. It may be excellent for the £200k local charity that has been doing the same thing for forty years.

What will struggle

In addition to email as a viable growth channel, other approaches are going to struggle in this new world of plenty (of asks). Things such as John Lewis-style TV ads that buy you a blip of awareness and not much else. The donor pyramid as a fundraising strategy, because the base of the pyramid is precisely where saturation and demographics hit hardest. Innovation without evidence. And the fundraising director whose entire toolkit was “spend more on acquisition” or “DM is dead” has not got long either.

DM engages, and will keep engaging. The fact is that shit DM doesn’t work. It never has. If your DM isn’t working, get a new agency – one that actually likes and understands old people, mid-value donors and legacies.

The sprints on AI, along with the dives, the safaris, the futures workshop, the five truths, the nine levers, the call to start with purpose. All of these are useful in the way a well-run workshop is useful. None of them is a strategy. The sector has produced an enormous amount of thoughtful, careful, values-led writing about AI adoption over the last three years, and a fair bit of it is going to look, in hindsight, like deckchair arrangement on a ship whose actual problem was a great big hunk of floating ice.

And let’s not forget, a chunk of the sector’s growth over the last twenty years was paid for by channel economics that have probably stopped working. A lot of charities have cost structures built on the assumption that those economics will continue. They will not. The reckoning is not five years away. And most fundraisers have not noticed because the lagging indicators are still moving in roughly the right direction.

So what stays?

People still want to give. They still want to belong to something larger than themselves. They still want to feel that the thing they gave to is real, that the people running it can be trusted, and that their money does something they would be proud of. None of this changes. As ever, people want to be a valued member of a valued group.

What changes is which channels can still carry that signal without getting drowned in noise, and which audiences are still around to receive it. The two key factors of digital saturation and a refocused demographic inversion, hit different channels for different reasons, and most fundraisers aren’t linking the two together. They are not separate problems. They are the same problem from two directions, and most of the sector’s behaviour has been lacking focus since well before Covid.

Here’s my honest synthesis. AI is here and it is going to be a big part of our lives. The genie is out of the bottle. And the best we can hope for is the approach of Anthropic is the one that wins.

But what does that actually mean in practice for individual giving fundraisers? Their jobs could be seriously under threat. If the Blackbaud offer is taken up en masse, then AI will be managing data, segmentation, targeting, writing and developing comms, thanking, reporting back, and building donor journeys with legacy fundraising at the end. Add in new product development, and what is left for the IG fundraiser to do? Blackbaud says there will be more time for strategy and human interaction. Which is a polite way of saying: how many current fundraisers actually have the top-end skills for that?

And then when we consider the position of the donor, they are going to be awash with asks. There used to be an approach to IG fundraising called “Goldilocks”. The more you ask, the more you got. Need more money? Add another appeal into the cycle. There lies the digital future as the digital commons are laid bare.

In short, individual giving is going to get harder, narrower, more relationship-dependent, and much more concentrated in the hands of charities that can do four specific things.

• Hold their existing donors as if their lives depend on it, because they do.

• Do excellent embodied human fundraising in places where the conversation can actually happen.

• Invest seriously in legacies when people are actively making their decisions about which charities will find a place in their will - that’s when people are in their seventies and eighties - not their forties.

• And convert the whole posture from “acquisition machine” to “community of meaning that people want to belong to.”

None of this appears on the standard sector AI opportunity list. None of it is what “start with your why” gets you to either.

So what does the work look like?

Go back to the things that will get increasingly get valuable in a world packed through of AI.

• Build connection.

• Thank well.

• Focus on specific or local, visible work.

In short, authenticity.

It may not win a Cannes Lion. But these are the things AI is structurally worst at faking, and they are the things older donors, the ones with the money, have been telling the sector they want for decades.

That is not a coincidence. The channels that are about to break are the ones the sector built when it stopped listening to its existing donors and started chasing automation and scale. The channels that are about to get more valuable are the ones it stopped paying attention to in roughly the same period. The donors who are about to matter most are the ones it has spent a decade trying to replace with Millennials.

Which is the loop back to the beginning of this piece. The fundraising sector has a serious case of FOMO. The agencies will keep selling the next big thing, because that is what they do and what many fundraisers want. The friendly LinkedIn posts will keep announcing it. AI will keep being the topic, then purpose will be the topic, then whatever comes next will be the topic. None of that is the work. Fundraising isn’t fashion, so let’s please stop treating it like it is.

Enshittification is endemic in the fundraising sector, and it has been brought about by people who should know better. AI is not going to fix this. Not without the sector seriously rethinking how it should be used. It is increasingly part of what broke things in the first place. We have handed over our human relationships to be commercialised and automated, because doing them properly is too expensive, too boring, or too hard.

That was the mistake. Everything that has destroyed trust, whether it is the rebrand or the next clever product to recruit younger donors, will not just turn out to have been a waste of time and money. It will prove to have been part of the problem.

Even the call to “start with a clearer purpose” is not a substitute for the work that comes next. That is what gets offered when nobody wants to face what the uncomfortable truth actually is.

Hold on to the donors you already have. Treat them like the most valuable resource you have, because they are. Thank them properly, and take it offline. Do better fundraising, in fewer places, for the people who actually want to hear from you. Patiently build toward the gifts that get read out in wills a decade from now. That is the work. Everything else is just looking busy.

The acid test is what you have already done. If you have spent the last few years chasing the next big thing, this is where it really starts to show.

One slide from this week’s CIOF Fundraising Convention is going to stay with me. The Blackbaud plenary, the same session that introduced the “Agents for Good” category, included a slide that read “the most valuable skill in the AI era is human judgement.”

Quite. Use yours.

PS. if you want more on this

If you are free on Tuesday lunchtime (9th of June) UK time, I’ll be speaking about enshittification on a short webinar with Fastmap. You can sign up here.

And for my Australian and New Zealand fundraising friends. I will be on tour at the end of July, working through all this – and loads of other stuff – in person. The dangers, the opportunities, and what to do about both, so you can set yourself up for what comes next. Get your ticket by clicking here.