From discussion to deployment: building a habitat recommendation tool in a day
The best design sprints aren't really about speed, they're about preparation meeting the right conditions. A hackathon with Public Realm Innovation Lab and Dark Matter Labs reminded me of that, and quietly changed how I approach early-stage product work
In just a day, we were able to put together a prototype of a Habitat Recommendation Tool that assists local authorities in making informed decisions on habitat creation and species selection in urban environments.
The tool unifies scattered data sources, enabling non-experts to make informed decisions and start urban habitat creation projects with confidence. It integrates geospatial data, policy layers, and ecological considerations to recommend the best plant species for specific urban sites.
In reality – this wasn’t just the outcome of ‘just a day’ – the hackathon was focused on outcomes, based on research, strategic direction and a pretty tight design brief.
My role was the translation of stakeholder needs into potential user journeys and product architecture, then developing the prototype spun up in Lovable.
Faciliated discussion to actionable output with NotebookLM
The real discovery for me was the approach taken to translate facilitated discussion into actionable output. I’ve since adapted a version of this approach for internal workshops at Agreena. The assembled team represented a broad range of perspectives (importantly we had a member who was the intended user of this type of application), and as such the conversation took many directions over the course of the morning.
By lunchtime, using the recorded transcript and prompt in NotebookLM, we had a PRD ready to take to prototyping. Perfection wasn’t the aim here, but by the afternoon we had a working prototype that we were able to share with the wider group and get feedback on.
Using AI to get more input into early stages of work
To me this felt like a great demonstration of how we can use AI to bring more input into design work at early stages. The lag or gap between inputs and outcomes is reduced to such an extent that feedback can be acted on in real time. The most valuable thing AI offers design isn't automation, it's compression. More voices, faster feedback, earlier in the process
The hard design work remains human: problem framing, team formation, and reconciling competing needs for societal, environmental, and economic benefit can't be automated away.
Find out more about the project and prototype here