Initiative overview
Public consultation is a legal requirement in planning, and public participation is critical - especially as planning issues often provoke strong local interest. However, local authorities often face an overwhelming volume of feedback, sometimes ranging from short notes to submissions of over 100 pages long.
Plan AI employs large language models to summarise and analyse these responses to public consultations conducted by local authority planning departments.
The tool was initially tested on three live planning consultations undertaken by Greater Cambridge Shared Planning relating to the Cambridge Biomedical Campus, Planning Obligations, and Health Impact Assessments.
The trial results were striking: PlanAI generated detailed summaries of every individual submission and a compendium report for all three consultations in just 16 minutes. By contrast, human planners took over 60 hours to log and summarise the 320 public submissions, including over 18 hours creating summaries – longer would have been required if human planners had been tasked with producing an overview analysis of the consultation exercise as Plan AI does. In terms of the quality of the summaries, the project team found that there were no noticeable differences between the AI summaries when compared to the human generated ones. The AI-generated reports included descriptive statistics and analytical insights, offering a clear, data-driven overview of the submissions’ content and themes.
The AI tool was developed by Professors Alex Lord, Alex Singleton and Dr Mark Green from the University of Liverpool’s Department of Geography & Planning working in partnership with the Planning Policy and Strategy team at Greater Cambridge Shared Planning. This close collaboration was an important part of the development of the AI tool in making sure that the reports that it generates were consistent with how the Councils currently report back to local communities and elected members. The next phase for the development of Plan AI will be working with five new pilot authorities selected by the national government Ministry for Housing Communities and Local Government.
Results, outcomes and impacts:
PlanAI has undergone rigorous testing, including an initial pilot using 300 responses from three consultations. It summarised all submissions in just 16 minutes, compared with 18.5 hours for human planners, demonstrating its potential to streamline plan making and reduce administrative burdens. The tool also has a minimal carbon footprint—roughly equivalent to sending two emails. Following GCSP’s successful use of PlanAI, the Ministry for Housing, Communities and Local Government has commissioned five additional pilots to assess how this purpose-built planning technology could be scaled nationally. Early results are highly encouraging, indicating that PlanAI could deliver a step change in planning practice.
