Google Cloud: AI Transforms UK Urban Planning

Google Cloud and the Automation of Local Councils in the UK

British ministries have begun deploying Google Cloud's generative artificial intelligence in municipal agencies to automate council planning operations. This initiative aims to alleviate the massive administrative burden associated with managing unstructured data, which hinders infrastructure development.

The UK central government has set an ambitious target: to build 1.5 million new homes by 2029. However, local planning authorities are facing administrative delays due to excessive documentation, significantly slowing down the development process.

To overcome these obstacles, the Ministry of Housing, Communities and Local Government (MHCLG) and the Department for Science, Innovation and Technology (DSIT) have implemented two machine learning tools. These tools are designed to accelerate the processing of municipal applications. At the Google Cloud Summit in London, officials announced the national rollout of the 'Extract' application and the progress of the 'Augmented Planning Decisions' (APD) prototype.

Lila Ibrahim, Head of AI Readiness at Google DeepMind, emphasized the importance of this initiative: “The UK has the opportunity to build the homes our communities need, but local councils are facing a mountain of paperwork. That’s why we are co-creating a sophisticated planning tool directly with councils to address real bottlenecks.”

She added that these tools will significantly reduce decision-making times, allowing planners to focus on the future to advance construction in the UK more quickly.

The Challenges of Housing Applications

Housing applications, which include common home modifications such as loft conversions or property extensions, account for nearly 70% of all planning applications submitted each year. Evaluating these standard submissions requires planning officers to spend hours cross-referencing regional policy documents, historical archives, and unstructured PDF files.

This repetitive evaluation process consumes administrative hours that could otherwise support major infrastructure and commercial developments. The deployment of automation aims to reduce decision times for applications by 50%.

Capabilities of Generative AI Tools

Engineers from the MHCLG and the government's Applied AI team, the Incubator for AI (i.AI), built the Extract tool in-house using Gemini base models. After trials in over 20 local planning authorities, administrators have expanded the application to all councils in England.

Extract analyzes unstructured data locked in old PDF archives, converting hundreds of pages of historical planning documentation into structured digital datasets in just minutes. Operational data from the trial phases indicate that the tool will eliminate approximately 255 hours of manual data entry per council each year. This reduction allows local authorities to reassign staff to complex evaluation tasks.

Integrating large language models into public sector workflows requires enterprise-level security environments. Local authorities handle sensitive civic records, necessitating strict risk management protocols to prevent data exposure.

The government has hosted the Gemini models on Google Cloud to establish a protected operational environment where data sovereignty is maintained. The cloud environment features active security controls to block malicious entries, including prompt injection attacks. This technical framework ensures that sensitive municipal data remains secure during testing and production cycles.

Features of the APD System

The APD system, on the other hand, acts as an analytical assistant for municipal planning officers by automating four main administrative tasks:

• The system consolidates incoming documentation by pre-processing data backlogs, flagging gaps in missing information, and extracting essential geographic data onto a unified user interface for review by officers.

• The software identifies relevant national and local zoning laws, assesses compliance margins, and adds precise policy citations for manual verification.

• The application analyzes public consultation letters, summarizing objections from stakeholders or historical legal precedents.

• The model generates initial drafts of final assessment reports, including technical reasoning and recommended approval conditions.

Protocols dictate that human planning officers retain final decision-making authority over each application. The software does not independently automate final approvals or rejections. Staff members review every line of text generated by the machine learning models, modifying the analytical reasoning before validating the report.

To maintain regulatory accountability, the APD prototype records its internal processing steps sequentially. This mechanism establishes an auditable thought chain, creating a trail for each processed application to support the officer's final determination.

Trials and Large-Scale Deployment

The development of the APD prototype relies on a collaborative framework linking public sector administrators with engineering teams from Google Cloud, Google DeepMind, and Faculty.

The alpha version is currently being tested live in three local authorities: the London Borough of Barnet, Dorset Council, and the London Borough of Camden. Testing in these different regional jurisdictions provides developers with varied municipal datasets to test the software against diverse local policies.

Central planners aim to complete the alpha phase and deploy the APD tool to over 300 English local authorities by 2027. Google Cloud provides the elastic computing infrastructure necessary to handle the thousands of simultaneous inference requests generated during daily operations.

Paul Maltby, Director of Public Services at Faculty, commented: “The English planning system is clogged. Planning officers are forced to spend half their time reviewing applications to convert a loft, putting on hold those for housing estates and warehouses.”

He added that the AI system, designed with planning officers, will eliminate the drudgery of reviewing simple planning applications so they can make quick decisions. This will allow officers to focus on major developments that matter, and importantly, enable families to improve 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, undertake a preliminary assessment, and draft the foundations of a report has the potential to save significant time for officers working on the administration of planning applications and direct that towards accelerating the decision-making process for residents. This will significantly contribute to achieving our growth objectives for housing construction in the borough.”

The coordination between the MHCLG, i.AI, Google DeepMind, and Faculty establishes a structured division of labor for enterprise software engineering. Public ministries define policy guidelines and statutory limits, while external technical partners design and deploy the underlying model architectures.

The successful integration of these systems demonstrates the feasibility of hosting advanced language models within a secure public cloud infrastructure to handle essential administrative workloads and modernize public service delivery.