Equity First: How AI Helped Us Move More Black-led Organisations Forward
At Do it Now Now, equity is at the heart of everything we do. When it comes to grant-making, this means ensuring that every organisation—especially those led by and serving underrepresented communities—gets a fair chance to access funding. In 2024, as we introduced AI into our screening process for the first time, we remained committed to making decisions that reflected this core value.
We piloted AI screening across five funding rounds: Innovate Now (Rounds 1-3) and Make It Big (Rounds 1 & 2). Our goal was to streamline our process, reducing the time spent on initial application reviews so that we could focus on deeper engagement with applicants. However, we knew that AI isn’t perfect, and relying on it without oversight could reinforce existing barriers rather than remove them.
To ensure we continued to act in the best interests of our community, we made an intentional choice: we trusted the AI when it said “yes” to an application but double-checked every “no.” Every organisation that was initially rejected by AI was manually reviewed by our team. We cross-checked details with Companies House and the Charity Commission and directly contacted organisations if their application details conflicted with official records.
Through this process, we discovered that 33% of applications initially rejected by AI were due to minor applicant errors. These were often simple mistakes—selecting the wrong geographic location (e.g., choosing "Kent" instead of "Greater London"), inputting an organisation name slightly differently than it appeared in official records, or accidentally misclassifying their work. Had we relied solely on AI, these organisations could have been unfairly disqualified from funding opportunities.
Instead of treating AI as an absolute decision-maker, we used it as a time-saving tool that allowed us to reinvest our efforts into supporting more organisations. By reducing time spent on early-stage screenings, we were able to move more organisations forward to our next phase: the Organisational Health Check and Due Diligence process. This phase involves interviews with applicants, deeper verification of submitted information, and risk assessment to ensure funding decisions are fair, informed, and impactful.
Our biggest takeaway is clear: AI should serve as a tool for equity, not a barrier to it. While it helped us make the screening process more efficient, human oversight was essential in ensuring fairness. This experience also reinforced the need for clearer guidance for applicants, helping them avoid common mistakes that could affect their eligibility.
Looking ahead, we plan to:
Provide better applicant support to help organisations avoid common form errors.
Refine our AI model to flag possible mistakes before an application is submitted.
Continue using a hybrid approach where AI supports decision-making, but people remain at the centre of the process.
We believe that technology should enhance accessibility, not create additional hurdles. AI can play a role in equitable funding, but only if it is designed and used with intentionality.
