2-Hour AI Hackathon: How 200 Consultants Became Copilot Builders
A 200-person AI hackathon producing 24 working agents in a single two-hour session.
Corporate AI Hackathon Case Study
Type: 2-hour internal AI hackathon (Microsoft Copilot)
Participants: ~200 (cross-company)
Output: 24 working agents built · 6 finalists
Key objectives: Hands-on AI adoption, Copilot up-skilling, deployable agent use cases
About FourCentric
Structure: A group of 6+ professional-services businesses
HQ: London
Industry: Supply Chain Consulting & Professional Services
FourCentric wanted to move its people from talking about AI to building with it. Rather than another training deck, they ran a high-energy, 200-person AI hackathon where cross-company teams built working Microsoft Copilot agents against real consulting challenges. MHD Innovation facilitated the day and synthesised the results. In a single morning, a largely first-time-Copilot room produced 24 working agents and a clear shortlist of tools worth taking into the business.
Hackathon Objectives
Adoption over theory: get hundreds of people hands-on with Copilot in one sitting, whatever their starting level.
Real consulting problems: anchor the build in three challenges drawn from day-to-day delivery, on realistic data.
A pipeline to deployment: surface the agents genuinely worth rolling out — not just the best pitch.
Process: Map → Hack → Deploy
FourCentric's leadership designed the event around three real-world challenges, each backed by a realistic data pack from Anvil Analytical.
MHD Innovation shaped the event structure and team briefs so teams could hit the ground running in a fast, high-energy session.
On the day, each of the 26 teams followed the same three-stage journey:
Map: Unpack the challenge. Teams went through the brief and data pack to understand the problem, what they had to work with and what a useful agent would need to do.
Hack: Build it in Copilot. Teams turned their idea into a working agent, testing and refining it with support from Anvil's technical mentors.
Deploy: Think beyond the demo. Teams defined how their agent could be used in real delivery and what should happen next, then pitched to a judging panel scoring feasibility and impact.
After the event, MHD Innovation reviewed each submission and synthesised the results into a debrief covering themes, the standout agents and a clear "take forward" recommendation.
Tangible Results
24 working agents built across 26 teams in a two-hour window
Near-total participation in a deliberately mixed-ability room. First-time Copilot users shipped working demos
A deployment shortlist the group is now progressing
Final pitches: a hackathon team demos their working Copilot agent.
Outcome Highlights
Teams built agents across three areas of day-to-day consulting work: project delivery, proposal generation, and supplier performance.
One lesson stood out: the strongest agents closed the "last mile" to a finished, usable deliverable. That's the line between an interesting demo and a tool people actually adopt.
From one morning to real momentum
Beyond the adoption goal met on the day, the event left FourCentric with a live internal AI community and a practical shortlist of tools to build on.
“It’s amazing what 200 people can build in two hours with the right structure and facilitation. Teams went from Copilot beginners to demoing working agents, and the energy in the room was incredible!”
Ready to get your teams building with AI in hours, not months?
MHD Innovation designs and facilitates high-energy AI hackathons, then turns the output into a clear plan for what to deploy next.
Ready to take the first step? Book a Call today
