We Hacked the Hackathon: Building AI Capability Through Culture

Not long after I joined Circle Medical to lead People and Culture, our CEO set a bold challenge: everyone in the company should be building AI capability into their work. We had already grown from a single AI hire into a full AI Tiger Team, and we wanted to bring the whole company along. So we teamed up across People, L&D, Clinical, Product, and AI to create our first internal AI Hackathon. It was not just a tech event. It was a culture builder.

Day one: designing the experience

We applied design thinking to the hackathon itself, pairing cross-functional collaboration with a Lunch and Learn on career growth and closing the day with a dinner focused on human connection. The goal was real collaboration between clinicians, product managers, and AI experts, with provider and patient perspectives kept front and center.

Day two: into the work

Our internal AI and Product experts ran a hands-on masterclass. We opened the hood, demoed live tools already in use, and let teams tackle real problems with AI, design thinking, and a clear lens on responsibility and risk. Our most ambitious prototype broke, and we welcomed it, because that is how learning happens. From whiteboards to wins and raw prototypes to real breakthroughs, it was a genuine moment of growth and connection. We went on to scale the model company-wide, and we kept learning as we went.

If you want to build your own

What to do:

• Set clear dimensions of focus. We structured ours around AI tool development, design thinking, and graduated levels of risk assessment and mitigation.
• Offer pre-work. We provided internal intros to key concepts, expert videos, and a few tasks and prompts so people arrived with baseline familiarity with our AI tools.
• Ground it in real work. Team members proposed areas where AI could improve their own roles, and those became the seeds for the brainstorms.
• Design for all skill levels. The programming intentionally supported everyone from the AI-curious to the AI-proficient.
• Measure impact. We ran a pre-session survey to gauge starting exposure and a follow-up survey to capture insights and track change.

What to avoid:

• Do not assume everyone finished the pre-work. Verifying follow-through was harder than we expected.
• Do not wait too long to schedule next steps. Momentum fades fast, so plan the leadership presentations and deeper integration before the event ends.

What worked best was strong engagement, genuine cross-functional collaboration, and a real sense of value and connection across the team. What we kept refining was pre-work completion and the structure of follow-through after the event.

One resource worth sharing: Anthropic's AI Fluency course is a free, high-quality primer that helps teams build a shared language around AI. It is an excellent addition to any pre-work curriculum, especially for non-technical team members.

How are you building AI capability inside your own organization?