Arthur.ai

A few colleagues and I built Arthur, a contextual AI learning tool, from scratch in a two-day hackathon. The idea came from an existing problem across the Learning & Development (L&D) industry: generic e-learning that is not role or context-specific. Research has shown that training becomes dramatically more effective if it reflects someone's actual role, colleagues and context.

Arthur uses AI and the Microsoft ecosystem (Graph, CoPilot, etc.) to make training specific to the person doing it, particularly personalising the content to the learner’s real day-to-day context, rather than a one-size-fits-all scenario. A cybersecurity module, for example, could reference a real colleague and a realistic project instead of a made-up name and generic red flags. The training structure and topics can be built once so it stays the same underneath, but everyone experiences a different, unique version of the content.

I acted as product owner and filled the design gap myself (there wasn't a dedicated designer on the team). On the final day, I pitched the idea to Microsoft and Nasstar, telling the story of why the problem mattered and how Arthur solved it. Arthur won the hackathon outright, and it's since evolved from a two-day prototype into an ongoing consultancy and training offering for L&D teams.