Government departments are rolling out Google Cloud’s generative AI tools across local councils to streamline planning operations.
Public sector bodies manage enormous amounts of unstructured data that slow down infrastructure projects. The UK government has set a goal to build 1.5 million new homes by 2029. Local planning authorities face administrative bottlenecks due to heavy paperwork, which delays these construction targets.
To tackle these challenges, the Ministry of Housing, Communities and Local Government (MHCLG) and the Department for Science, Innovation and Technology (DSIT) scaled up two machine learning tools designed to speed up local government processing. At the Google Cloud Summit London, officials announced the nationwide rollout of the ‘Extract’ tool and the advancement of the ‘Augmented Planning Decisions’ (APD) prototype.
Lila Ibrahim, Chief AI Readiness Officer at Google DeepMind, said: “The UK has a chance to build the homes our communities need, but local councils are buried under paperwork. That’s why we’re working directly with councils to develop a smart planning tool that tackles real-world bottlenecks.
“This will dramatically shorten decision times, allowing planners to focus on the future and help Britain build faster.”
Householder applications – covering routine home modifications like loft conversions or extensions – make up nearly 70 percent of all planning applications submitted each year. Assessing these standard applications manually forces planning officers to spend hours checking regional policy documents, historical records, and unstructured PDF files.
This repetitive review process eats up administrative time that could otherwise go toward major infrastructure and commercial projects. The introduction of automation aims to rebalance this workload, targeting a 50 percent reduction in application decision times.
Core capabilities of the Google Cloud generative AI tools
Engineers at MHCLG and the government’s applied AI team, the Incubator for AI (i.AI), developed the Extract tool in-house using Gemini foundation models. After testing in more than 20 local planning authorities, officials expanded the tool to every council in England.
Extract processes unstructured data trapped in legacy PDF records, turning hundreds of pages of historical planning documents into structured digital datasets in minutes. Trial data suggests the tool will save around 255 hours of manual data entry per council each year. This frees up local authorities to redirect staff to more complex assessment work.
Bringing large language models into public sector workflows demands enterprise-grade security. Local authorities handle sensitive civic records, requiring strict risk management measures to prevent data leaks.
The government ran the Gemini models on Google Cloud to create a secure operating environment where data sovereignty is preserved. The cloud platform includes active security controls to block malicious inputs, such as prompt injection attacks. This setup ensures sensitive municipal data stays protected during both testing and live production runs.
The APD system, meanwhile, serves as an analytical assistant for local planning officers by automating four key administrative tasks:
- The system organises incoming documents by clearing data backlogs, highlighting missing information, and pulling out key geographical site details onto a single dashboard for officer review.
- The software finds relevant national and local planning regulations, evaluates compliance levels, and adds specific policy references for manual checking.
- The tool reads public consultation responses, summarising community objections or past legal precedents.
- The model produces initial drafts of final assessment reports, including the technical reasoning and suggested approval conditions.
Under established protocols, human planning officers keep final decision-making power over every application. The software does not independently approve or reject applications. Staff members review every line of text produced by the machine learning models, adjusting the analysis before signing off on the report.
To ensure regulatory accountability, the APD prototype logs its internal processing steps in sequence. This creates an auditable chain of thought, providing a verification trail for every application processed to support the officer’s final decision.
Local council planning trials and scaling timelines
The APD prototype is being developed through a partnership between public sector officials and engineering teams from Google Cloud, Google DeepMind, and Faculty.
The alpha version is being tested live in three local authorities: the London Borough of Barnet, Dorset Council, and the London Borough of Camden. Testing across these different regions gives developers access to varied municipal datasets to evaluate the software against diverse local policies.
Central planners aim to finish the alpha phase and roll out the APD tool to all 300-plus English local authorities by 2027. Google Cloud supplies the scalable computing infrastructure needed to handle the thousands of simultaneous queries generated during daily operations.
Paul Maltby, Director of Public Services at Faculty, commented: “The English planning system is overwhelmed. Planning officers are forced to spend half their time reviewing applications for attic conversions, while applications for housing estates and warehouses sit on hold.
“Developed alongside planning officers, our AI system will remove the tedious work from reviewing straightforward planning applications so they can make faster decisions. It will allow planning officers to concentrate on the major developments that matter, and importantly, let families upgrade their homes without months of delay and uncertainty.”
Naisha Polaine, Executive Director for Growth at Barnet Council, added: “The tool’s ability to gather relevant information, carry out a preliminary assessment, and draft the foundations of a report has the potential to save considerable officer time spent on planning application administration and redirect this toward speeding up the decision-making process for residents. In turn, this will play a major role in meeting our house building growth targets in the borough.”
The collaboration between MHCLG, i.AI, Google DeepMind, and Faculty sets up a clear division of responsibilities for enterprise software development. Public ministries set the policy guidelines and legal boundaries, while external technical partners build and deploy the underlying model systems.
The successful deployment of these systems shows that advanced language models can be hosted within a secure public cloud environment to handle core administrative tasks and modernise public service delivery.
See also: EU publishes its AI content labelling playbook ahead of the AI Act’s August deadline
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information.
AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.
