Skip to main content

Weeknotes

A collection of regular communication about work in progress on the alpha for Extract.

Subscribe to the feed to get the latest content in your RSS reader. Learn more about feeds.

  1. Second round of research

    Rapid turnaround of a prototype led to research with five users, uncovering unexpected insights – plus some wins. It’s all about that test-and-learn approach, adapting the plan for what users actually need.

  2. Making improvements for users

    We’re focusing on what users actually need – from simple data uploads to tools which match their existing processes. By working with AI’s limitations and listening closely to users, we’re uncovering the dynamics of human-AI collaboration.

  3. Early insights from users

    Early analysis indicates that the core concept is well understood and shows strong potential to deliver effective, highly valued solutions. The team also explored the latest AI models for improved map analysis.

  4. Benchmarking begins

    Weeknote for Extract Alpha covering the deployment of 0.1, first LPA testing, and model experiments with Gemini 3 and Segment Anything 3.

  5. First research session booked

    We’ve reached out to 14 local planning authorities and booked a research session with the first responders. The plans for the first prototype are finalised, and it’s being built ready for testing next week. Plus how to align AI’s actions with user intent.

  6. Forming the research plan, and extracting trees

    Jordy showcased a new feature for tree extraction in TPOs, while Gavin presented georeferencing tech improvements. Rob outlined our research roadmap, and the team planned key features for testing.

  7. “A strong week, a solid 9”

    Finding code we can re-use, sharing knowledge in a big collaboration session, and onboarding our user researcher. We’re building up steam.

  8. The unit of delivery is the team

    This week we kicked off the alpha, held our first planning session, and started looking at risky assumptions.