Planning --- The War of the AI Tools?

Planning delays have long been one of the biggest frustrations in England’s housing system.  So it is not surprising that the government is very active in investing in AI tools to aid the planning process.

Government Use of AI

Extract

The government is rolling out  it’s first large AI tool designed to drag historic planning records out of paper archives and into the digital age. Ministers say it could help speed up decisions across England by turning old planning files into usable data in a fraction of the time currently required.

At present, councils often spend one to two hours processing a single document, whether it is a paper map, a PDF or a handwritten note. Extract is intended to cut through that bottleneck. Built with large language models and computer vision tools, it can pull text from legacy documents, analyse maps and images, and convert decades-old planning material into structured digital records within minutes. Ministers have framed that as a practical fix for one of the planning system’s least glamorous but most persistent delays.

Key Features include:

• Text extraction: pulls key details such as dates, locations and planning decisions from historic files.
• Image and map analysis: interprets scanned plans, blurry maps and handwritten annotations that would otherwise need manual checking.
• Georeferencing: aligns historic boundaries and planning records with modern digital mapping systems.

Early trials suggest the gains could be significant. In pilot testing, Extract digitised planning records in roughly three minutes each, compared with the one to two hours officers would usually spend doing the work by hand.

Pilot Trials and Deployment

The tool has been tested with councils including Hillingdon, Westminster, Nuneaton & Bedworth and Exeter. The government says it wants Extract available to all councils by Spring 2026, and officials argue that its wider value lies not just in faster validation but in creating the clean, standardised data needed for the next generation of digital planning systems, including tools such as PlanX.

Benefits

• Faster decisions: by reducing document-processing delays, the tool could help applications move through the system more quickly.
• Time and cost savings: ministers say it could free thousands of officer hours now lost to repetitive manual work.
• Better planning data: structured records would give councils, developers and policymakers a more reliable basis for decision-making.
• Support for housing targets: the wider aim is to remove friction from the system as the government pushes to deliver 1.5 million new homes.

Augmented Planning Decision Tool

The next phase is even more ambitious. Councils are preparing to test a Google-built tool designed not just to digitise records, but to help assess planning applications themselves. The pitch is straightforward: cut backlogs, speed up routine decisions and give overstretched departments more analytical firepower.

The system is not meant to replace planners, at least not on paper. Instead, it is being developed to sift through digitised planning data, local policy and supporting documents before producing a recommendation on whether an application should be approved or refused. Officials say the final call will remain with council officers, but the ambition is clear: routine applications could be processed far faster, and eventually close to instantly in the simplest cases.

Supporters see an obvious upside: less pressure on under-resourced planning teams and quicker movement on the applications clogging the system. Critics are less convinced. Faster decisions, they argue, are not automatically better ones, particularly if transparency, professional judgement and public confidence are weakened along the way. The central question is whether AI can make planning more efficient without making it seem more opaque.

Behind that sits a larger question. If councils start using AI to assess applications, developers, campaigners and consultants are unlikely to stand still. They, too, may turn to competing models, rival datasets and machine-assisted arguments about what policy, precedent and local impact really mean. In that sense, this is about more than speeding up paperwork. It may mark the start of a new contest over how planning decisions are framed, challenged and ultimately made.

Citizens Use of AI in Planning

Until recently, lodging a credible objection to a retail park, brownfield redevelopment or housing scheme usually meant hiring a planning consultant at a cost that could run into the thousands. AI has sharply lowered that barrier, turning what was once specialist work into a service that can now be bought in minutes.

Services such as Objector.ai and PlanningObjection.com are part of a small but fast-growing market in AI-assisted planning objections aimed directly at residents rather than professionals.

Objector.ai says it can generate “policy-backed objections in minutes” for £45 on a full planning application, while also offering a £249 crowdfunded option for larger outline schemes. PlanningObjection.com, meanwhile, markets its service as a way to produce “persuasive, policy-centred objection letters” at a fraction of the cost of a planning consultant.

Beyond dedicated platforms, there is growing evidence that residents are also using general-purpose tools such as ChatGPT to produce large volumes of bespoke objections to individual applications, each tailored just enough to avoid being dismissed as a duplicate.

Geoff Keal, chief executive of TerraQuest, which runs the Planning Portal in partnership with government and says it covers roughly 95 per cent of planning applications in England, argues that councils are already feeling the effect. New AI tools, he says, are enabling objectors “to be able to provide better objection documents, much wider and much broader, which is slowing the system down, because obviously those things need to be dealt with in the right way... It’s certainly what we’re seeing local authorities suffer from.”

For councils already under strain, that creates a practical problem. Officers cannot simply ignore submissions that cite the National Planning Policy Framework, local plans and case law, even when they suspect a chatbot has done much of the drafting. Every objection still has to be logged, assessed and, where material, answered.

That matters because the wider system is already under pressure. The Home Builders Federation says the number of sites granted permission in England fell to its lowest rolling annual total since the series began two decades ago, with only 8,200 sites approved in the 12 months to June 2025.

Digital democracy

Supporters of public-facing AI tools say this is not sabotage but a form of planning democracy catching up. For years, developers and well-funded applicants have been able to deploy consultants, lawyers and policy specialists with ease, while residents have often struggled to frame their objections in the technical language that planning committees expect.

Hannah George, co-founder of Objector.ai, says the service was built to help residents produce “high-quality, evidence-based objections” while cutting down on invalid, repetitive or purely emotional submissions. She says the platform discourages users from relying on generic chatbots to mass-produce letters and instead screens applications free of charge before advising whether there are valid planning grounds to object at all.

It is also a mistake to treat AI as the original cause of the backlog. Pressure on the planning system long predates chatbots, and ministers have already framed opposition to development as a wider political problem. Rachel Reeves has pledged to confront what she sees as a culture of obstruction, while a broader planning overhaul led by Angela Rayner is intended to push more decisions through more quickly. AI is arriving in a system that was already struggling to cope.

Attitude of the Planning Inspectorate

The Planning Inspectorate, which handles appeals against council planning refusals, has moved to get ahead of the issue. In guidance updated in February 2026, it said applicants and objectors can use AI in casework evidence, but only if they do so transparently and responsibly. Anyone using tools such as ChatGPT or Microsoft Copilot to draft or substantially rewrite submissions is expected to say so. The warning is blunt: if parties fail to disclose significant AI use, they risk weakening the credibility of their case.

How AI is changing the Planning Landscape

First, the era of frictionless local opposition is over. Developers can no longer treat sophisticated objections as an exception; AI-assisted resistance is fast becoming a standard feature of the planning process.

Second, early and credible community engagement is likely to matter more than ever. The developers who talk to residents before an application is filed, rather than after opposition hardens, will be better placed to avoid costly delays, especially if they lack large in-house public affairs teams.

Third, scrutiny of AI itself is set to increase. Applicants should expect councils and inspectors to ask sharper questions about how these tools are being used, by objectors and developers alike.

Britain’s planning system was under strain long before AI arrived. But cheap, capable tools on the objector’s side of the process are raising the stakes. They do not create the underlying dysfunction, but they do make the case for reform more urgent and the consequences of delay more costly for the businesses waiting on schemes that have yet to win approval.