Founderland supports women of colour founders in Europe and the UK who’ve faced obstacles in their business journeys because of their ethnicity and race. Only 0.5% of VC funding goes to women of colour founders. Founderland’s investor-readiness programme, Compass, helps change that, but it can only take so many people.
So the question we were given at R/GA was: could technology help prepare more founders each year for the moment they pitch to a VC?
Meet PAM
PAM (the Pitch AI Mentor) is a safe space to practise. A founder runs through a five-minute pitch and PAM gives instant, personal feedback on their delivery, their slides and the content of the story they’re telling.
There’s also a simulated live pitch mode with virtual investors. Founders choose whether they want a supportive panel or a properly tough one, and they can see sample answers to the questions they struggled with.
What happened
PAM launched in beta to a quarter of the Founderland community and its extended network, and was expected to support more than 2,000 women of colour entrepreneurs in Europe in its first year. Fast Company covered the launch.
“In our accelerator, we focus on early-stage entrepreneurs who typically haven’t had a chance to meet an investor yet. Pitching is a key skill they need, but it’s tough without practice or access to real investors. That’s where PAM comes in – it’s like a practice ground for these founders, giving them a fair shot at getting their ideas out there.”
Stephanie von Behr, co-founder, Founderland
Testing showed founders found the feedback really helped them improve their pitches, and gave them the confidence to go and do it for real. PAM went on to win awards, at a time when AI products were still rare in agencies.
AI written by AI
When I could finally talk about it, I wrote that it was AI for good: something meaningful and useful that wouldn’t have been feasible without the latest technology, both in the product itself and in how we put it together.
The real-time feedback is powered by an LLM, but the front-end and back-end codebases were also built with a lot of help from LLMs. It was AI written by AI. Designing and architecting a complex back-end system behind a simple, friendly interface was a real team effort, and it meant rewriting the rulebook on how we collaborated and built things. It let us move very quickly at a consistent code quality, and it opened my eyes to a new way of working: AI-assisted development is how I’ve built ever since.