There are more than 10,000 grant-making trusts and foundations in the UK, and only a handful will ever fund work like yours. Identifying them means reading eligibility rules, cross-referencing them against how your charity actually operates, and ranking them by fit, long before you write a single application.
I built AiDonor to do that work automatically.
What aiDonor does
AiDonor is an AI-powered donor matching and outreach tool. It reads every UK funder, understands each one’s funding criteria, and returns a ranked shortlist of your ten best matches. Each match comes with a fit score, the evidence behind it, and the strongest angle to lead with. What used to take a fundraiser weeks now takes minutes.
A multi-agent system, not a chatbot
AiDonor runs as six AI agents, each doing the job a member of a fundraising team would do. One builds a structured profile of your charity. One enriches and maintains a database of every UK funder. One reads each funder’s criteria and scores it against your profile. Others handle eligibility, outreach and deadlines. It doesn’t match keywords or rely on basic filtering. It reads a funder’s own rules and reasons about fit the way an experienced fundraiser would, across thousands of funders at once.
Transparency, engineered into the core
Most AI tools are a black box. AiDonor is built to be the opposite. Every decision is explainable. Every match shows its evidence, its provenance, and how confident the model is. When a funder has ruled a charity out, AiDonor shows the exact rule in the funder’s own words and lets you contest it. Every decision is logged and auditable, and a person approves anything before it leaves the system. For a sector accountable to funders and regulators, that transparency isn’t optional.
Built with Afghan College for Girls
I built AiDonor in partnership with Afghan College for Girls, a charity teaching 500 girls in Afghanistan who are banned from school. Their model is unusual: salaries paid in the UK, impact delivered overseas, no capital spent abroad. Most international funders exclude exactly that. AiDonor surfaced a trust near the top of the shortlist because one rule in its criteria matched precisely, a match a human scanning funder names would have missed entirely.
This one sits outside the Power Platform work I usually write about, but the way it was scoped, built and handed over is the same. If you have a job that a spreadsheet or an inbox is quietly failing at, tell me what it does and I will tell you whether it is worth building something for.