TL;DR
More than 3,000 small plots with housing potential have been found across Lewisham using an AI mapping tool, in a pilot that MHCLG has now written up as a case study. The project came out of the PropTech Innovation Fund’s 2024 round.
Named the Small Sites AI Finder, the tool was developed alongside RCKa, a practice working in architecture and urban design, and set to work across the borough’s land. It replaced a process leaning heavily on individual officers’ knowledge of their patch plus manual review — workable for spot checks, poor for producing a consistent borough-wide picture or an evidence base a planning policy can rest on.
Officers retained the decision. The case study is explicit that AI-generated candidates still needed professional review before any site could be taken seriously, and MHCLG framed the value as strengthening evidence gathering rather than substituting for planning judgment. Officers could also probe what each plot looked like and what limited it, ahead of committing to detailed assessment. Alongside the mapping work, the council examined planning application data to gauge how far its design guidance on small sites had influenced delivery.
The contrast with other councils is instructive on where the sector actually is. Solihull Council announced this week that supplier ICS.AI will lead phase one of the council’s AI Transformation Discovery work — 24 weeks spanning adult social care, services for children, the economy and infrastructure brief, and public health. The intended output is roughly 200 candidate uses, of which 50 would be checked against finance, plus a roadmap. Councillor Dave Pinwell framed it as assembling evidence to inform later decisions.
One authority has a working tool and a number; the other is still scoping which problems are worth solving. Both are reasonable positions, and the gap between them is roughly where most of English local government sits.
Looking Forward
The Lewisham pilot appears beside other case studies covering site identification, planning assessment and local plan development, giving councils something more useful than a vendor pitch: a documented outcome from a comparable authority. With housing delivery under sustained political pressure, expect the small-sites approach to be tested elsewhere before the underlying evidence about site quality is fully in.