The grant application has been the backbone of charitable funding for decades. But it may not survive the next five years. At Craigmyle, we think the sector needs to start thinking about what comes next before the question becomes urgent.
There is a problem quietly spreading through the charitable funding world, and most people involved can sense it even if they haven’t named it yet.
Funders are receiving more applications. Many of them are well-written, clearly structured and professionally presented. They arrive from organisations that the funder has never heard of, organisations with modest track records, organisations that would previously have lacked the capacity or confidence to submit at all. On the surface, this looks like democratisation. In practice, it is creating a new kind of noise.
Large language models (LLMs), the technology behind tools like ChatGPT and Claude, have made it easy to produce polished grant applications. Not exceptional ones, but competent ones. The friction that once filtered out the uncommitted or the under-resourced has largely disappeared. And so funders, already stretched, are drowning in a rising tide of submissions that are increasingly difficult to distinguish from one another.
The instinctive response, to use AI to screen on the other side, is understandable but troubling. If we build a system in which machines write applications for machines to read, with human judgement reserved only for the final shortlist, we haven’t solved the problem of funding finding purpose. We’ve just automated the theatre of it.
This matters for Craigmyle, and it matters for every organisation and funder we work with. Understanding what is actually breaking, and what might replace it, is not a technical question. It is a question about the relationship between funders and the charities they support, and about whether that relationship is fit for purpose in a world where AI is becoming embedded in all of our systems and processes.
What the application form was actually doing
The grant application form was never just a document. It was an information-gathering mechanism designed for a world of scarcity, where funders had limited means to learn about organisations, and organisations had limited capacity to tell their story. The form existed to discover identity (who is this organisation?), evidence (what have they done?), intent (what do they want to do next?) and fit (does this align with what we fund?).
That bundle of functions, compressed into a deadline-driven document, made a kind of sense when writing the document was itself a signal. Completing a well-structured application required time, sector knowledge and organisational coherence. The process was, in effect, a proxy for capability and quality.
That proxy has been broken. The signal no longer works. And because funders are now uncertain what the signal means, the entire mechanism is under stress.
This is not a problem that can be solved by adding more questions, shortening word counts or moving to Expression of Interest stages, though all of those are reasonable adaptations. The deeper issue is that the form-based model bundles things together that probably need to be separated: the discovery of organisations, the assessment of credibility and fit, the understanding of ideas and intent, and the decision to fund.
A different relationship is possible
The question worth asking is not “how do we fix the application form?” but “what would we build if we were designing the funder-charity relationship from scratch today?”
There is a school of thought, and a growing body of practice, that says the answer is purely relational. No forms, no rounds, no competitive processes. Just funders getting to know organisations over time, building trust and funding on the basis of these relationships. Some funders have moved in this direction and there is much to commend it.
But purely relational funding carries its own risks. Relationships are not equally accessible. The organisations doing the most important work in the most under-served communities are often the furthest from funder networks. A world in which funding flows through relationships, without any balancing infrastructure, is a world that systematically advantages the already-connected. This is not a hypothetical risk. It is a documented pattern, which clearly embeds more inequality into the sector, not less, as so many of us are working to achieve.
The more interesting question is how to build the infrastructure that enables relationships to form on more equitable terms, and how different funders, at different scales and with different resources, can begin to experiment with that now.
What different funders can actually do
There is no single answer here. The right approach depends on a trust’s size, staffing, thematic focus and appetite for change. But the space of options is wider than most funders currently explore.
For smaller trusts with limited capacity, the most accessible shift is probably moving from periodic competitive rounds to a continuous open register. Organisations can submit a lightweight profile at any time, describing who they are and what they do. The trust then makes proactive contact when they see a fit, rather than waiting for organisations to apply at the next funding round. This decouples the administrative burden from the funding timeline. It requires basic CRM discipline but not significant staffing. And it makes the trust accessible to organisations that lack the capacity to monitor funding calendars and respond to deadlines.
For mid-sized trusts with programme staff, proactive discovery becomes viable. Rather than publishing a call and processing what arrives, programme staff develop genuine knowledge of the landscape, through field visits, peer networks and systematic desk research, and then approach organisations directly. The trust becomes a scout rather than a gatekeeper. Done well, this produces far richer understanding than any written submission. Done badly, it replicates the same network biases as purely relational funding. The discipline required is a deliberately broad scouting methodology, not just funding the organisations the programme officer already knows.
For larger foundations with significant resources, participatory and community-led grantmaking approaches become possible. Panels of people with lived experience of the issues the trust addresses make the funding decisions, within parameters set by the foundation. This shifts the accountability for quality from the funder’s judgement to the community’s. The evidence from funders who have tried this seriously is encouraging, though it requires investment in supporting panellists and resisting the urge to override their decisions when they diverge from institutional instinct.
Across all sizes, there is a strong case for portfolio-style funding with staged commitment, small initial grants to a wider cohort, with follow-on funding based on what is actually learned rather than what was promised in a business plan. This requires trustees to be comfortable with genuine uncertainty, which is a cultural shift as much as a governance one. But it produces far more honest relationships between funders and funded organisations, and far better information about what is actually working.
The AI dimension
None of this is separable from the broader question of what AI means for how organisations tell their story and how funders understand what they’re reading.
AI is not arriving in the charity sector. It is already here, in the writing tools that organisations use, in the screening tools that funders are beginning to deploy, in the data analysis and reporting systems that sit behind the numbers in grant applications. The question is not whether to engage with it but how to engage with it thoughtfully.
The risk we are most concerned about at Craigmyle is not the obvious one, that AI will produce fraudulent applications or mislead funders about what organisations do. The greater risk is subtler: that AI accelerates a trend towards ‘style over substance’, at the expense of deep, authentic understanding. That the signals funders have historically used to assess quality, clarity of thinking, quality of evidence, organisational coherence, become easier to simulate, and in becoming easier to simulate, become less useful as signals.
If that happens, the funders who will maintain genuine insight into the sector are those who invest in relationships, site visits and longitudinal knowledge of organisations, the kinds of understanding that cannot be generated by a LLM working from a charity’s website and annual report.
This is, in a way, an argument for going back to basics. The best funders have always known that the application form tells you less than you think. AI makes that truth impossible to ignore.
The question for trustees and programme staff
The honest question for every charitable trust is this: if your current approach to discovering and assessing organisations is primarily form-based, what happens when the form becomes an unreliable signal? What is your plan B?
Some funders are already asking this question seriously. Most are not, yet. The organisations and funders that think it through in advance, before the dysfunction becomes acute, will be better placed to build something that genuinely serves the sector.
At Craigmyle, we’re not offering a blueprint. We don’t think one exists or that it should. But we are committed to exploring these questions with the funders and organisations we work with and to helping the sector think carefully about the relationship between resource and purpose before the tools that have shaped it for decades stop doing the job.
The application form has had a good run. What comes next deserves serious thought.
Craigmyle is one of the UK’s leading fundraising consultancies, working with charities and funders across strategic development, organisational change and fundraising campaigns. If you’d like to explore what these questions mean for your organisation, we’d welcome the conversation. Get in touch at first@craigmyle.org.uk or call 01582 762441 to arrange a call.
