AI and impact: the new frontier for funders

This question came into focus at two recent NPC Funder drop-ins with AIConfident and The National Lottery Community Fund. NPC’s Funder drop-ins are informal, closed sessions designed to give funders space to explore emerging issues together, share practice openly, and test thinking with peers under Chatham House rule.  

Understanding the AI journey: four levels of change

One way to make sense of this shift is to think about four levels of AI-driven change, as shared at our recent funder drop-in by Tim Cook, founder of AIConfident.

1. Individual productivity: how staff use AI tools to support day-to-day work, from drafting documents to summarising information more efficiently.
2. Process automation: where AI streamlines internal workflows. For funders, this might include summarising applications or supporting analysis, with a clear red line: decision-making should not be left to AI.
3. Service delivery: where AI becomes directly relevant to grantmakers. Charities are beginning to use AI as part of what beneficiaries’ experience. For example, in sectors such as mental health, carefully designed tools can expand access and reduce stigma, particularly for those less likely to engage with traditional services. But poorly designed tools can cause harm. In 2023, the National Eating Disorders Association suspended an AI chatbot after it gave harmful weight-loss advice that could worsen eating disorders.
4. Societal change: AI is reshaping how people access information, how misinformation spreads, how work is structured, and how vulnerable communities experience risk. Funders need to understand this context. It shapes both the needs charities respond to and the risks they face.

As Tom Ilube, the Chair of The King’s Trust, put it at last year’s Ignites event: ‘This stuff is really, really coming… we can’t be left behind and let the tech companies dictate to us, we need to step into it.’

What this looks like in practice: lessons from The National Lottery Community Fund

Kirsten Nicholson, Innovation Manager at The National Lottery Community Fund, showed what this looks like in practice.

TNLCF’s AI work began in 2024 and has focused on three strands: internal experimentation, organisational culture, and sector-wide convening. A key message from Kirsten was that responsible AI adoption is as much about people and culture as it is about technology.

Internally, the Fund has run controlled experiments using tools like Microsoft Copilot. For example, a summarisation tool created first drafts of project summaries in seconds, with humans firmly in the loop to review and approve outputs.

Another pilot explored whether an AI agent could triage routine HR queries, freeing up staff time to focus on more complex or sensitive issues. Importantly, Kirsten emphasised that AI is not always the right solution: sometimes the real problem is access to information, not the absence of automation.

Alongside this, the Fund invested in culture: training, clear policies, shared principles, and peer learning.

Openness was critical, moving from ‘don’t tell the boss I used AI’ to transparent conversations about what works, what doesn’t, and where risks remain.

Beyond readiness: a role for funders

This points to a broader challenge for philanthropy.

Charities will apply for funding to deliver AI-supported services. They will also use AI to write applications. Funders may use AI internally. The question is not whether this will happen, but whether funders are prepared to engage with it responsibly and proactively.

Funders need to move beyond preparing to assess AI-enabled proposals and towards supporting organisations to use AI well, while actively managing risks.

This includes investing in capacity, ethical frameworks, and shared learning, not just technology.

What we’re doing at NPC

One area of interest is the role of AI in Theory of Change development.

As charities face increasing pressure to articulate complex impact, AI may support sense-making and iteration, if used thoughtfully, with human judgement at the centre. NPC is exploring this further. Please get in touch if you are interested in funding this.

Internationally, some funders are already moving at scale. In 2025, the MacArthur Foundation joined a coalition to launch Humanity AI, a $500m initiative to build a people-centred future for AI.

There is no equivalent coordinated effort in the UK. This is both a gap and an opportunity.

UK funders will need to build confidence internally. But there is also a case for proactively shaping how AI enhances capacity and service delivery across the sector.

The question is no longer just, ‘are we ready to review AI-enabled proposals?’, it is:

What role do we want UK philanthropy to play in shaping how AI is used for public good?