Faculty is taking a greater role in bringing artificial intelligence into Britain's public services, as its planning technology, PlanAssist, moves into an expanded trial across 15 English councils. Developed alongside Google DeepMind, Google Cloud and the UK Government, the system has already reduced the time planning officers spend on key administrative tasks by 58 per cent, offering a potential route to faster decisions on routine housing applications.

The project aims to reduce the time taken to process straightforward residential applications, such as home extensions and loft conversions, from the current standard of eight weeks to four. For Faculty, the initiative demonstrates how AI can be applied to practical problems within government, helping public-sector teams manage workloads without removing the expertise needed to make important decisions.

The initial phase involved close collaboration with planning officers at three local authorities: Barnet, Dorset and Camden. Faculty worked directly with these teams to design and test the system around their everyday requirements. Early benchmarking showed that key administrative tasks took 58 per cent less time, effectively halving the processing time associated with routine householder applications.

The improvement addresses a persistent challenge for local planning departments. Officers must check whether a property falls within a conservation area or flood-risk zone, assess proposals against local and national planning policies, and gather evidence before recommending an outcome. These checks can consume considerable time, even for relatively minor home improvements.

PlanAssist is designed to take much of this groundwork off their desks. It brings planning applications, relevant policies and site constraints into one place, helping officers identify the information they need without repeatedly searching through separate documents and records.

Crucially, the system does not make planning decisions independently. It prepares an evidence-based initial assessment, while the planning officer retains responsibility for reviewing the findings and deciding whether to accept, amend or reject the recommendations. Faculty's approach places human oversight at the centre of the technology, recognising that planning decisions often involve local considerations that require professional judgement.

Feedback from officers during the initial trial also influenced how the platform was developed. Early versions followed a rigid, step-by-step process, presenting recommendations in a fixed sequence. Officers found this restrictive, prompting a shift towards a more flexible dashboard. The updated design allows users to move between aerial maps, planning constraints and other relevant information in the order that best suits each application.

The system also assembles an evidence base covering material considerations that could influence a decision. This includes potential impacts on neighbouring properties, such as the loss of daylight caused by a proposed extension. By bringing these details together, PlanAssist aims to make the assessment process more organised and transparent. It also explains how it has applied relevant policies, allowing officers to examine its reasoning and challenge findings where necessary.

The next stage will extend testing to 12 additional councils over a 12-month beta programme, bringing the total number of participating authorities to 15. The wider group has been selected to represent different geographical areas, demographics, planning policies and working practices. This will help establish whether the technology can perform consistently beyond the three authorities involved in its initial development.

The beta phase will also move the project closer to everyday council operations. During the initial trial, information was entered manually through static forms. Faculty and its partners will now work towards connecting PlanAssist directly to the existing case-management systems of two councils, allowing it to process live applications as they arrive.

For Faculty, the longer-term opportunity lies in helping public services use AI to reduce administrative pressure and improve operational efficiency. If routine applications require less manual work, planning officers could devote more time to complex or contentious cases, where local knowledge and careful judgement remain essential.

The expanded programme will test whether the system is safe, reliable and suitable for wider adoption. Faculty will continue working with councils to refine the technology before it moves towards a permanent operational model. With the Government aiming to make the tool available to councils across England by the end of 2027, the results could help determine how AI is adopted across local planning services.

‍