PolicyBot

TL;DR

While working at a London-based real estate development firm, I observed the inefficiencies in navigating the planning phase of the development cycle. This often involved significant costs and delays due to complex, borough-specific regulations.

Inspired by tools like LandInsight, I realized there was an opportunity to use technology to streamline the planning process. This vision led to the creation of PolicyBot, an AI-driven planning assistant tailored to London’s real estate market.

problem

The planning phase is the most complex and critical step in real estate development. High-quality planning is essential, as it determines whether a project will proceed smoothly or face costly delays and overruns. London’s planning environment, governed by the London Plan and 33 independent borough councils, presents a unique challenge. Developers must navigate layers of regulations, many of which vary significantly between boroughs.

This complexity often results in planning refusals. In 2023, 12 of London’s boroughs rejected more than 50% of major housing proposals due to misalignments with local guidelines. Traditional planning advisors helped bridge this gap but at a significant cost, with initial project reviews averaging £1,000 each. Developers like ourselves needed a faster, more affordable way to align their projects with regulations.

solution

I developed PolicyBot to address these inefficiencies. Starting with a prototype focused on Westminster, I used Replit and the GPT-4 API to build an AI-powered assistant that could interpret planning guidelines and provide instant, accurate insights. Over time, I scaled the platform to include planning regulations for all London boroughs.

Later on, to enhance functionality and user experience, I integrated Danswer, an open-source AI tool, and optimized the platform for scalability with AWS hosting. PolicyBot aggregates data from multiple sources, such as the London Plan, local council guidelines, and key legislation, enabling developers to quickly verify the feasibility of their projects.

result

Since its launch, PolicyBot has proven to be a game-changer for developers and the real estate community. Internally, PolicyBot helped us eliminate nearly £3,000 in monthly advisor fees, allowing resources to be redirected to other critical areas.

Beyond cost savings, PolicyBot improves the quality of planning applications by helping developers align with local guidelines from the outset. This has the potential to reduce planning refusals and expedite housing approvals, directly supporting London’s goal of meeting its growing housing demand.

Today, PolicyBot is live and available for free, with over 50 users actively testing the platform. By gathering user feedback, I aim to refine the tool further, ensuring it continues to deliver value to London’s real estate community while showcasing my ability to conceptualize, build, and scale impactful products.