AI4C

The work behind the timeline

AI4C is an example of how I prefer to lead: recognise the potential in a good idea, give capable people room to develop it, and put enough support and governance around it to last. Its creation and growth were collective. My contribution was not to claim operational ownership, but to provide sustained institutional sponsorship, constructive challenge and a route from an experimental exchange to an established programme.

Turning two grants into shared infrastructure

The starting point was the 2018 WCIT Charity IT Award. CALM and Missing People each received £300,000 to develop services using artificial intelligence and machine learning. The awards were substantial, but the more transferable idea was the learning exchange created alongside them.

Technology projects rarely proceed exactly as planned. Suppliers change, data proves less usable than expected, internal capability varies and apparently simple ideas expose difficult legal, ethical and operational questions. If the experience remained inside two recipient organisations, much of its wider value would be lost. AI4C created a way for charities to compare experience while projects were still under way, including the difficulties and failures that conventional case studies often remove.

I returned to the Charity Board in January 2019 and became Chair in June. The awards and the exchange had already been created. My responsibility was to help ensure that the third element of the award, the shared learning, did not disappear when the original projects or their working group came to an end.

Creating trust, not another sales channel

AI4C’s value depended on the character of the space it created. It was free to participating charities, supplier-neutral, technology-agnostic and sales-free. Meetings operated under the Chatham House Rule. Members were expected to share not only successes but also uncertainty, delay, risk and lessons learnt.

Communicating the 'bad bits' was really important. A charity considering AI may need to know that it is not ready, that its data requires attention first, or that a proposed product is poorly suited to its operating model. Impartial advice can prevent wasted money as well as unlock worthwhile innovation. The exchange therefore offered something different from a conference programme or a collection of polished supplier presentations: a continuing community in which people could confidently ask unfinished questions.

The early output was substantial. During its first three years, AI4C recorded 20 workshops, seven events, more than 140 lessons learnt, 31 presentations, five charity chatbot demonstrations, 25 guides and two practical toolkits. Members could return to that material through a shared online library and continue discussions between meetings.

Putting governance around a growing programme

An informal initiative can lose momentum when its originating project ends or a key volunteer moves on. In 2020, I raised the need for AI4C to have a durable reporting mechanism once the time-limited IT Awards working group ceased to exist. I worked with the Clerk and Charity Coordinator on the structure, and the Board appointed Gary Moore as the link trustee.

When that link later became vacant, I asked Serena Chaudhry Pandey to take responsibility for reporting AI4C’s progress to the Board. This gave the programme a clear line of accountability without drawing its day-to-day work into trustee meetings. Its terms of reference were subsequently refreshed and expanded to cover not-for-profit organisations as well as registered charities.

The Charity also supported the less visible infrastructure around the programme: coordination, communications, a dedicated website, branding, online working space and a designated budget. I was directly involved in the 2020 website and identity work and in ensuring that the Board continued to consider AI4C’s resourcing and future rather than treating it as a temporary accompaniment to the original grants.

Backing the people who made it work

The programme’s progress depended on people prepared to give time, knowledge and leadership. Maxine Ricketts made a tremendous contribution to establishing AI4C and creating its early network. When she stepped down, Paul Excell took over as AI4C Chair and expanded its activity, reach and practical support. Serena Chaudhry Pandey provided the reporting connection to the Board. WCIT members, external experts and participating charities supplied presentations, advice, pro bono help and honest accounts of their own experience.

My role was to support and challenge those leaders, remove institutional obstacles and encourage ambition.

It also explains why succession was important. I supported Paul as he took over, welcomed the momentum he created and made sure that programme leadership had an effective route back to the Charity Board. The aim was a strong programme, not dependence on the Chair of its parent charity.

Growing with the questions charities were asking

AI4C changed as artificial intelligence changed. Its early work drew heavily on the practical experience of CALM and Missing People, including chatbots, data quality, procurement and the business case for AI. Later programmes addressed large language models, generative AI, security, regulation, ethics, accreditation, fundraising, service design, AI agents and the legal issues facing charities.

The format changed too. During the pandemic, the community created material that members could use asynchronously and moved activity online. Later, quarterly hybrid workshops were supplemented by virtual coffee conversations, ad hoc webinars, direct introductions and problem-solving between meetings. In 2024, the AI4C Chair began developing guidance aimed specifically at trustees and charity leaders and represented the non-profit perspective in wider professional and policy discussions.

This responsiveness was important because interest in AI did not automatically create preparedness. The 2024 Charity Digital Skills Report found that almost two thirds of charities regarded AI as relevant, while fewer than a quarter felt ready to respond to its opportunities and challenges. AI4C offered a practical bridge between those two positions.

From a small exchange to a national community

The clearest measure of growth is participation. Fourteen charities were involved in 2019. The figure passed 25 in 2021, approached 30 in 2022, exceeded 35 in 2023, reached 67 at the end of 2024 and 82 by the end of 2025.

 The community included organisations working in mental health, disability, cancer, poverty, safeguarding, homelessness, advice, research and family support. Large national charities and much smaller organisations could bring different capabilities to the same table. WCIT members and other specialists contributed expertise, while charities contributed the operational realities that made the discussion useful.

By December 2025, the Board heard that the network was no longer mainly listening. Members were collaborating and experimenting. That was an important change. The exchange had moved beyond explaining AI to charities and was helping charities learn by doing, together.