
Whether your organisation is just starting to experiment or you want generative AI to become a dependable part of delivery, I help teams use it on real work, with enough care that leaders can stand behind it. The approach is practical and hands-on, shaped in public-sector and regulated environments as well as product and research organisations.
A focused engagement for one team that is ready to use generative AI on the work in front of them. Together we pick something real — a report, a prototype, a manual process that hurts — and use tools such as Cursor, Copilot, Gemini, or ChatGPT until your people can carry on without me in the room. You leave with a shared way of working, a few checks that fit your context, and the confidence to keep going on Monday.
Good practice spreads when people have a place to show their work. I’ve helped a cross-government AI community of thousands through events, newsletters, and open guidance, with facilitators who keep the conversation alive. We can start much smaller. A regular meetup, a short newsletter, and a few people who care is enough for teams to learn from each other and build on what’s already working.
Sometimes the fastest way to a decision is something people can click. I join your team as a forward-deployed engineer and build a thin, real tool: a catalogue of what you’re already using, a way to explore overlapping guidance, a dashboard, or the first version of a service. Recent work includes the Public Sector AI Tools catalogue and an AI Guidance Explorer. The first version of the explorer came together in under two weeks. We stay on lightweight stacks — Vue, Python, Google Apps Script, AWS — which helps when procurement would otherwise slow you down. I wrote up how the catalogue came together.
Leaders need a story they can trust. I help capture what your organisation already knows — meetings, guidance, scattered documents — and turn it into reporting and decision support. Using data automation and generative AI, I’ve helped cut ministerial-level report creation from more than two weeks to under five days. We can also shape a lightweight way to test AI products, so confidence and adoption grow together. This fits the public sector and other regulated settings, and any organisation where the cost of getting it wrong is high.
These can be combined, shortened, or wrapped into a longer contract. AI work often sits inside a wider change: interim delivery and programme leadership, or OKR coaching when you also want clearer outcomes. See my CV for the longer story, then let's talk about what would actually help.