PlanAssist, an AI tool developed by Google DeepMind with the UK government, Google Cloud and Faculty, reduced the time planning officers spend assessing householder applications by up to 58% in early pilots that began in June. The system automates routine checks so officers can focus on matters that require professional judgement, while retaining human approval for every decision.
Who built it and what is the aim?
Google DeepMind developed PlanAssist in collaboration with the UK government, Google Cloud and Faculty. The tool is intended to automate repetitive administrative tasks in planning workflows so that officers spend less time on routine processing and more on complex cases.
PlanAssist is part of a broader partnership between Google DeepMind and the UK government exploring how AI can transform planning and other public services. The system is built using the Gemini model.
How PlanAssist supports officers
- Householder applications can account for up to 70% of planning officers’ workloads, despite often requiring straightforward checks.
- PlanAssist validates submitted documents and architectural drawings to ensure required information is present, checks applications against heritage constraints (for example tree preservation orders) and local and national planning policies, and reviews whether objections or issues have previously been raised by local residents.
- The tool produces fully cited draft reports that planning officers can edit before making a final decision.
In pilots across Barnet, Camden and Dorset since June, PlanAssist more than halved the average officer assessment time per application. Additionally, over 70% of the reports generated by the tool—covering findings such as compliance with local planning rules or completeness of documentation—were approved by officers without edits.
From pilot to scale
Following the early tests, PlanAssist is entering a 12-month Beta phase that will process live applications. Twelve additional local planning authorities will be onboarded in three cohorts, increasing the total number of participating councils to 15. These councils include metropolitan and rural districts, enabling testing across diverse communities.
Subject to successful testing, the government plans to make the tool available to all councils in England in 2027. The rollout plan maintains that every assessment will be reviewed and approved by a qualified planning officer before any decision is issued.
Why this matters
Planning delays carry economic and personal costs: they can slow progress toward the UK government’s target of building 1.5 million new homes by 2029 and extend household waits for approvals on projects such as loft conversions, conservatories or extensions. By taking on time-consuming administrative tasks while keeping final decisions with human officers, PlanAssist demonstrates how AI can reduce backlogs and speed up approvals.
Experience gathered during the Beta and the broader roll-out will determine whether and how the system is deployed more widely across England.
