Agents who Argue
A Facilitated Workshop for: University of Birmingham
The situation and brief
The University of Birmingham asked Treehouse to help run a one-day, in-person workshop called “Agents Who Argue,” hosted by the Centre for AI in Government. The aim was to bring together technology experts and researchers to explore how AI agents could help with deliberation, negotiation, and decision-making. The workshop was designed to uncover real-world challenges, match them with new technical ideas, and spark partnerships for future projects - all in a single day.
The Treehouse solution and activity
Treehouse designed and delivered a fast-paced, hands-on workshop using its sandpit approach. Stakeholders shared their real-world problems, and teams worked together to brainstorm ideas, use empathy mapping, and create visual summaries with drawings and keywords. Groups developed “How Might We…?” questions and pitched early project ideas, making sure everyone could contribute and build on each other’s thinking.
After the event, Treehouse wrote up a summary for the host, capturing all the information and ideas for future use.
The outcome or impact
The sandpit sparked new research topics and teams, with several project pocket pitches shared. Participants found it easy to connect, share ideas, and work together. The write-up gave the host a clear record to build on for future AI research and collaboration. Feedback was very positive, with participants describing the event as energising and a catalyst for fresh thinking and momentum in AI research.
Client feedback
“🌞 Despite the heat, the Agents Who Argue workshop was a fantastic success!
On Tuesday, we welcomed a brilliant group of researchers and professionals from across academia and industry to the University of Birmingham for a deep dive into LLM-based Multi-Agent Systems.
Expertly facilitated by Becky Evans and Gill Bates from Treehouse, the day focused on designing deliberative interactions between generative agents to support robust scenario planning, particularly in complex decision-making contexts.
We also tackled the challenges and opportunities of applying this methodology in areas like justice and social services, where trust, nuance and accountability are key.
We’re excited to build on the momentum of the day and continue the conversations that emerged.”
