Google DeepMind says PlanAssist, a planning tool built with Gemini, brought the average time planning officers spent on householder assessment tasks down by 58% in early pilot testing at Barnet, Camden and Dorset councils. The tool is to be made available to an additional 12 local planning authorities across England, the company wrote on Saturday.
That takes the councils involved from three to 15, five times the pilot by our arithmetic. More than 70% of the reports the tool generated in the pilot were approved by officers with no edits required, according to the post.
Short explainer · Planners’ time -58%
A vertical news Short summarising this story. Figures come from Google DeepMind’s blog post of 31 October; our arithmetic is labelled.
Business Engineer Course · The FDE Playbook
An evergreen Business Engineer course on forward-deployed engineering, how AI gets put to work inside organisations. It teaches the general idea only and does not discuss PlanAssist.
The key insight:Google DeepMind says its Gemini-built PlanAssist took 58% off planning officers’ time on householder assessment tasks in a three-council pilot, while officers keep the final say on every application.
PlanAssist was developed by Google DeepMind with the UK government, Google Cloud and Faculty. The post says it works in four steps. It validates application documents and architectural drawings, checks the application against heritage constraints such as tree preservation orders and against local and national planning policies, checks whether objections have been raised, for example by local residents, and drafts “fully cited reports” that officers can edit.
The company says the tool is “designed to support human judgement” and that planning officers retain the final say on every application. Routine householder applications, such as loft conversions, conservatories and extensions, “account for up to 70% of planning officers’ workloads”, the post says.
Google DeepMind’s PlanAssist pilot at Barnet, Camden and Dorset: officer time on householder assessment tasks down 58%, and 12 more councils joining a 12-month Beta. Company figures; the 25-minute and 5x figures are FourWeekMBA arithmetic.
What the Pilot Showed
The pilot has run since June across the three councils. The 58% figure is an average cut in officer time on householder assessment tasks; the company’s post on X phrased it as “up to 58%”. By our arithmetic, a 58% cut means an hour of that work would take about 25 minutes.
When Google DeepMind announced the prototype in June, it said the goal was to help officers cut application decision times by 50%. The October figure measures something narrower: officer time on assessment tasks, not the time to a decision. The posts read do not give decision times.
The Business Lever: Public-Sector Workflows
PlanAssist now moves into a 12-month Beta phase on live applications. The 12 councils will be onboarded in three cohorts, which by our arithmetic averages four per cohort; the post does not give cohort sizes. It says the councils include metropolitan and rural districts.
Subject to successful testing, the government plans for the tool to be made available to all councils in England in 2027, with every assessment reviewed and approved by a qualified planning officer before a decision is made, according to the post.
The company ties the work to the UK government’s goal of building 1.5 million new homes by 2029. It describes PlanAssist as part of Google DeepMind’s broader partnership with the UK government on AI in planning and other public services. The June post said an earlier Gemini-built tool, Extract, had been made available to every council in England.
The Structural Read
As we read the post, the tool is aimed at the routine end of planning work: householder applications, which the company says account for up to 70% of officers’ workloads.
As we read it, the design keeps the decision with the officer: the tool validates, checks policy and drafts a cited report, and the officer edits and decides.
As we read it, the next test is scale: 15 councils on live applications for 12 months, before the government decides on a rollout to all councils in England in 2027.
Google DeepMind blog, 31 Oct 2026
“Designed to support human judgement, the tool speeds up routine checks while ensuring planning officers retain the final say on every application.”
Three Implications
FOR COUNCILS The post says more than 70% of the reports the tool generated in the pilot were approved by officers with no edits required.
FOR THE GOVERNMENT The post ties the tool to the UK government’s goal of building 1.5 million new homes by 2029.
FOR GOOGLE The post describes PlanAssist as part of Google DeepMind’s broader partnership with the UK government on AI in public services.
The framework says “Judgment becomes the highest-value layer of work.”
As we read it, PlanAssist is built the same way: the tool does the checks and the drafting, and the planning officer keeps the decision. The framework is the owner’s argument and does not discuss PlanAssist or its figures.
What Is Not Established
The posts read do not say what councils or the government pay for PlanAssist, how many applications the pilot covered, or how the 58% was measured. The figures are Google DeepMind’s own, and the posts read do not point to an independent evaluation.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
Google DeepMind says its Gemini-built PlanAssist cut the time UK planning officers spent on householder assessment tasks by 58% in a three-council pilot. Twelve more councils join a 12-month Beta, and the government plans a rollout to all councils in England in 2027, subject to testing.
A note on sourcing. All facts and figures come from Google DeepMind’s blog posts of 31 October and June 2026 and its post on X, read on 31 October; the pilot results are the company’s own. We did not contact Google DeepMind or the councils named.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
