The starting point
A team might begin with a broad frustration such as:
“It takes too long to find the information I need.”
Global insurance company
AI Week · Innovation workshop design
Turning business challenges into Agentic AI ideas
I helped design a global insurance company’s AI Week workshop experience and prepare facilitators across seven local business units. Those facilitators guided participating teams in developing 25+ Agentic AI concepts, which became inputs to subsequent agent development by the company’s AI team.
PROGRAMME SCALE & OUTPUT
- 5
- Markets
- 7
- Local business units
- 25
- Teams
- 100+
- Participants
- 25+
- Output: Agentic AI concepts
My contribution
- Co-developed the workshop flow and activities to give teams a practical route from business challenges to concepts they could explain and discuss.
- Helped shape participant-facing content—the opening narrative, innovation framing, Agentic AI inspiration and facilitator handoffs—to give participants context for each activity without overwhelming them with technical detail.
- Helped prepare facilitator guidance and supported two training sessions and one dry run, so facilitators across local business units had a shared understanding of the activities and how to adapt them.
My role focused on designing and enabling the experience within the wider team. Local facilitators guided the participating teams, who developed the concepts.
The challenge
The hackathon needed to help people with different perspectives collaborate and turn early observations into concepts that others could understand.
The design challenge was to provide enough structure to keep teams moving, while leaving room for exploration and giving facilitators a usable approach across markets.
The key design decision
Problem first, technology second
Day 1 focused on problem framing: teams used Rose, Thorn, Bud, affinity mapping and How Might We statements to identify friction, cluster observations and select an opportunity.
Day 2 introduced Agentic AI exploration: inspiration and the Creative Matrix supported idea generation, followed by comparison, concept posters and rapid presentations.
The sequence was intended to keep ideas connected to business needs and reduce the risk of starting with a technology in search of a problem. The Impact–Difficulty Matrix provided a way to compare and prioritise ideas using teams’ assessments; it did not establish technical feasibility or business value.
From activities to a shareable concept
- 01
Surface friction
Capture observations about what works, what is difficult and what could change.
Rose / Thorn / Bud
- 02
Identify patterns
Group related observations into themes for discussion.
Affinity mapping
- 03
Frame the opportunity
Express a selected challenge as a focused question.
How Might We
- 04
Explore possibilities
Generate alternative ways Agentic AI might address the opportunity.
Creative Matrix
- 05
Compare ideas
Discuss anticipated impact and difficulty to select a concept to develop.
Impact–Difficulty Matrix
- 06
Communicate
Give the concept a clear structure for presentation and discussion.
Concept poster + rapid presentation
Worked example — From vague frustration to a prioritised AI concept
ILLUSTRATIVE EXAMPLE · NOT A REPORTED TEAM OUTCOME
The risk
Jumping directly to a solution could lead to something generic such as:
“Let’s build an AI chatbot.”
The reframing
The workshop was designed to push the team to ask:
- Who experiences the problem?
- Where does the friction occur?
- What information is fragmented or difficult to access?
- What decisions are being delayed?
- What still requires human judgement?
01Discover
Surface examples of the current friction.
02Define
Cluster the observations and frame a clearer opportunity.
03Develop
Explore different ways AI might help retrieve, summarise, recommend or move work forward.
04Prioritise
Compare concepts using an Impact–Difficulty lens.
05Communicate
Turn the selected concept into a clear poster and rapid presentation.
The point was not to produce a perfect solution in two days. It was to move teams from a vague pain point to a concept grounded in a real business need.
Preparing facilitators across markets
I helped prepare facilitators to explain the activities, manage the pace and respond to different team dynamics. The guidance was designed to make the experience understandable and repeatable across local business units, while leaving room for facilitator judgement.
- 2
- Facilitator training sessions
- 1
- Dry run
The guidance translated the design into practical facilitation
- Purpose and instructions: what each activity was for and how to explain it clearly.
- Time and participation: how to manage the pace and encourage balanced contributions.
- Probing and adaptation: when to challenge thinking and how to respond when teams stalled or moved too quickly.
- Contingency options: “safety nets” when an activity did not progress as expected.
The aim was to support consistent delivery without scripting every interaction.
What the programme delivered
Participating teams developed 25+ Agentic AI concepts, using concept posters and rapid presentations to articulate their proposals. The company’s AI team subsequently used concepts from the hackathon to develop actual agents, extending the programme’s contribution beyond the workshop into practical AI development.
My contribution connected workshop design, participant materials and facilitator preparation into a process that supported teams in developing and communicating those concepts. Participating teams created the concepts; the company’s AI team carried out the subsequent agent development.
What I took away
This project reinforced that workshop design extends beyond choosing activities. It also means designing the instructions, transitions and facilitator support that help people navigate the experience.
It brought together my work in Design Thinking, AI innovation and facilitator enablement, with a focus on making a shared process usable across different teams.