The Last Social Mile: How AI Is Finally Reaching the Charity Sector
By Jake O’Gorman CF
Jake is Director of AI at 10Pearls, and author of AI and the Future of Doing Good: How charities are turning technology into impact and what gets in the way, drawn from nearly 700 hours of conversations with over 275 charity leaders across the UK and the USA. Download for free.
Ask almost any charity leader how the last few years have felt, and you’ll hear some version of the same answer: harder. There has perhaps never been an easy time to run a charity, but right now has a particular kind of difficulty to it.
Between 2024 and 2025, giving in the UK fell by around 10%, from £15.4bn to £14bn. There are roughly six million fewer donors than there were a decade ago, and last year alone an estimated 2.8 million people cancelled a regular gift. The picture isn’t much brighter elsewhere in the world.
Add to this a cruel piece of timing: as funding tightens, demand for services climbs. It’s a squeeze that is becoming all too familiar to many leaders. And the problem runs deeper than money. Trust, the currency every charity ultimately runs on, is wearing thin: almost 20% of people who no longer give say it’s because they don’t trust charities to spend their money wisely.
It’s little wonder, then, that conversations about AI are gaining momentum across the sector. For organisations that are stretched and asked, always, to do more with less, the promise that technology might help them do exactly that is a compelling one. And adoption is already happening. A 2026 study by Hanover Research found that 47% of nonprofit professionals already save between one and three hours a week using AI tools.
A changing landscape for giving
There is another reason this matters, and it has nothing to do with internal efficiency. AI is now the biggest philanthropic advisor in the world, with more and more people turning to it for advice on where to give. Even a routine Google search now answers with a “zero-click” summary. It’s perhaps no surprise, then, that so many charities are seeing their website traffic drop, with little visibility as to whether this new form of search is working to their benefit or their detriment.
Many charities are quickly realising that if they are invisible to these systems, or worse, misrepresented by them, they are becoming increasingly invisible to donors. Adapting to AI, then, isn’t only about adopting new tools inside the organisation, but also remaining visible to an increasingly AI-oriented world.
Tools that never quite fit
Technology, and AI especially, demands a great deal of thought, and it is rarely as simple as the pitch suggests. AI is offered up like a side of fries with every purchase, and surface-level solutions rarely result in anything more than surface-level impact. Add to that, charities don’t have the disposable income of the private sector, nor the ability to underwrite large capital expenditures on the promise of a return that may never arrive.
It doesn’t help that the nonprofit experience of technology has mostly been one of making do, of getting by with tools that were never really designed for them. It’s no secret that charities have long been an afterthought in the product roadmap, too small a market to shape it and too specialised to be served well by the off-the-shelf version. There has been little incentive for anyone to go the last social mile, to close the gap between what the software does and what the work actually requires. And so charities have learned to bend generic systems to unusual needs, to build workarounds on top of workarounds, and to live with the daily friction of technology that doesn’t quite do the job it’s supposed to.
There is also a great deal to be desired in the way the private sector itself is doing AI. It is possible, on almost any given day, to find fresh examples of AI behaving badly, of tools that fail the people they were meant to serve. “Move fast and break things” is not a philosophy that works well where vulnerable people are concerned. When the cost of getting it wrong is borne by a survivor, a patient, or a child, “we’ll fix it in the next release” is not good enough.
The market starts to catch up
Which is why it is genuinely encouraging that the large providers are starting to move closer to the charity market, offering not just more nonprofit-friendly pricing but, in some cases, purpose-built versions of their technology.
Salesforce is the clearest example. Back in 2008, Salesforce released the Nonprofit Starter Pack, which would later become the Nonprofit Success Pack (NPSP): a version of its platform pre-configured around the way charities actually work, with fundraising, donor relationships and grant management built in rather than bolted on. Paired with its Power of Us programme, which gives eligible nonprofits their first ten licences for free, it marked a shift in posture: from treating charities as a smaller kind of corporate customer to designing for them as a sector in their own right.
And it was never Salesforce’s achievement alone. NPSP was shaped and sustained by a motivated, talented community of nonprofit developers: the people who understood the work from the inside and gave their time to make the technology fit it. It is a model of what going the last social mile can look like, and a sign of where the rest of the market may finally be heading.
From building it yourself to switching it on
Salesforce’s platform for the sector has since moved on from NPSP to Nonprofit Cloud, and the latest chapter in that story is Agentforce Nonprofit, which builds nonprofit-specific agents directly into the Salesforce environment. Take, for example, the prospect research agent, which helps prepare a fundraiser for a meeting in minutes rather than hours of research.
Features like this mark an important shift for charities. For years, “using AI” meant needing expertise across a whole range of areas, from wiring up the data to heavily customising the tools, which for many was a high bar to reach. Tools like Agentforce don’t remove the need for that expertise entirely, but they move much of the heavy lifting into the platform itself, so that engaging with AI starts to look less like a construction project and more like turning on a capability you already own.
Getting the foundations right
This is something I’ve seen first-hand at 10Pearls, where I lead our AI offering and work with a great many nonprofits and charities. We are, amongst other things, a Salesforce systems integrator. Much of what we’re asked to do is help organisations bring these new capabilities to bear. Just as often, though, the real work is everything that has to be in place before AI even enters the picture. Because the uncomfortable truth is that most AI problems are really data problems, and most data problems are really organisational ones.
Our work with The Citizens Foundation (TCF) is a good illustration. TCF is Pakistan’s largest education charity, serving more than 320,000 students across over 2,250 schools, and yet it ran on disconnected systems with no single view of its own impact.
We helped unify donor information from four separate regional fundraising operations into a single, trusted Salesforce view, replacing fragmented, siloed records with one connected platform. That foundation now supports more informed decision-making, more streamlined operations, and fundraising that can actually scale. None of it is particularly shiny or glamorous, and often little of it is AI in the headline sense, but it is exactly the groundwork that makes responsible AI possible later.
Technology in service of the mission
Because in the end, most nonprofits don’t usually need more technology. What charities need instead are systems that give time back to their teams, strengthen their relationships with donors and volunteers, and support growth they can sustain. The real value of AI won’t be measured in hours saved or tools deployed, but in whether it frees people to do more of what only people can do: build trust, show up for their communities, and turn scarce resources into lasting impact. That is the last social mile in action, and for the first time in a long while, charities have a real chance to benefit from it.
How 10Pearls Helps Nonprofits Drive Greater Impact
10Pearls helps nonprofits turn Salesforce and AI into practical tools for greater mission impact. By combining deep Salesforce expertise with AI, data, cloud, and digital engineering capabilities, we help organizations streamline fundraising, strengthen volunteer engagement, modernize service delivery, and build secure, AI-ready data foundations.
Whether optimizing existing Salesforce investments or designing enterprise-scale transformation initiatives, we partner with nonprofits to reduce operational complexity, improve decision-making, and create the capacity needed to serve more communities.
References
