The simplest place to begin is language.
Imagine if an applicant could write their response in Urdu, Polish, Bengali or Arabic — because that is the language in which they think most clearly and express themselves most honestly. The system could translate it accurately into English for assessment, preserving meaning and nuance.
Or imagine if the application questions themselves could be translated instantly into an applicant's first language. Guidance notes could be explained in plain speech rather than dense policy text.
For many community leaders, English is functional but not expressive. They can run sessions, support families and manage volunteers, but when faced with formal written English they begin to doubt themselves.
Language should not be the gatekeeper of impact.
Embedded AI Grant Assistants
Funders could go further and build AI assistants directly into their application portals.
Not tools that write bids for applicants — but tools that support understanding.
An embedded assistant could explain what is meant by "demonstrate measurable outcomes." It could offer examples of how small volunteer led groups have described their impact. It could break complex questions into simpler steps. It could help someone structure their thoughts without changing the substance of what they are saying.
This would particularly benefit:
- First time applicants
- Volunteer led organisations
- People without formal education
- Leaders for whom English is not a first language
- Those who have never encountered funding terminology before
It would reduce the invisible advantage of already knowing the rules.
Other Ways AI Could Build Equity
- AI could help applicants develop realistic budgets, guiding them through common cost categories and helping them avoid common mistakes
- It could provide audio versions of application questions for people with dyslexia or low literacy
- It could gently flag when jargon has overtaken clarity, encouraging plain English over institutional tone
- It could help funders analyse who is not applying, identifying geographic, socio-economic or demographic blind spots and prompting targeted outreach
- It could allow conversational, chat based application processes rather than overwhelming multi-page forms
None of this lowers standards.
It removes friction.
A Different Framing
If funders genuinely want to reach working class communities, racially diverse neighbourhoods and informal grassroots groups, then equity must be designed into the process.
Encouraging responsible AI use, offering translation, embedding guidance tools and simplifying early stage applications all send a powerful message.
You belong here.
The mum on the estate should not lose out because she does not instinctively use institutional language. The retired youth worker should not be excluded because he does not know how to frame outcomes in technical terms.
If we believe in community led change, then we should welcome tools that help community leaders navigate systems that were not originally built with them in mind.
At a time when technology makes that possible, choosing not to do so would be the real missed opportunity.
Equity must be designed in from the start — not as an afterthought.
The tools exist. The question is whether we are willing to use them.