The most valuable AI is the kind that understands the problem before it tries to solve it
The decisions that shape our built environment, from planning approvals to Section 106 agreements to building regulations, are among the hardest to get right. One AI platform, four products, the full planning lifecycle.
See our first product →The problem we solve
The built environment runs on decisions that are too slow, too manual and too exposed to error
From minor householder extensions to major developments with complex Section 106 obligations, the processes that shape every structure around us still depend on manual review, fragmented information and overstretched expertise.
Planning approvals take weeks.
Whether it is a single rear extension or a 200-unit housing scheme, applications pass through multiple rounds of manual review. Councils, architects and applicants all wait while documents are checked against national and local policy by hand.
Regulations are deeply layered.
National planning policy, local plan designations, conservation constraints, Section 106 agreements, CIL obligations, building regulations. Every project, minor or major, sits inside a web of overlapping rules that general-purpose tools cannot navigate.
Generic AI is not enough.
Models trained on everything understand nothing deeply enough to be trusted with planning decisions or building regulations. The built environment needs AI engineered from the ground up for the domain.
The market
A system that touches everyone who builds, extends or renovates
From homeowners planning a kitchen extension to developers navigating Section 106 agreements on major housing schemes, the planning and building regulations system affects hundreds of thousands of people every year. We are starting in the UK, the market we know best, before bringing our B2C products to homeowners and self-builders internationally.
Across England alone. Each one requiring validation against national policy, local plans and site-specific constraints.
Extensions, loft conversions, new builds. Householder applications make up half of all planning decisions. The other half spans minor and major developments, each with their own complexity.
Each with their own local plan, conservation areas and supplementary guidance that must be checked alongside national policy.
Architect drawings, planning application fees, building regulations submissions. The cost of navigating the system falls on everyone who builds.
Consumers
Save time and money. Confidence before you submit. Launching UK-first, then internationally.
- Homeowners planning extensions
- Self-builders and renovators
- Landlords and property investors
Businesses
Reduce rework. Increase throughput. De-risk compliance across minor and major developments.
- Architects and planning consultants
- Developers modelling S106 obligations
- Housing associations
Public Sector
Faster decisions. Fairer outcomes. Do more with less.
- Local planning authorities
- Council digital transformation teams
- Building control bodies
Our approach
We go deep into the built environment. We understand planning policy, building regulations, Section 106 obligations and design compliance before we write a line of code.
We serve both sides of the planning system: the applicants who create and submit, and the authorities who validate and decide. Proprietary models trained on planning decisions, regulatory frameworks and architectural documents, validated against real outcomes and deployed where the decisions actually happen.
The result is a platform that has evolved from a single product into an end-to-end AI planning system. Four products, one intelligence layer, powered by our PlanCraft engine. UK-first, with international expansion planned for our consumer-facing tools.
Products
One platform. Four products. The full planning lifecycle
From application creation through validation, post-permission delivery and building regulations compliance. One shared intelligence layer powers every product.
Planning Studio
End-to-end platform for creating complete, submission-ready planning applications. Powered by PlanCraft, our proprietary drawing engine.
Register interest →Coming SoonPlanning AI
AI-powered validation and assessment for UK planning applications. Covers householder, minor and major developments.
Visit gpetoplanning.ai →LiveS106 Intelligence
AI platform for Section 106 agreement negotiation and drafting. Model contributions, timelines and triggers upfront.
Register interest →MHCLG-funded PilotBuilding Regs AI
AI engine for building regulations validation. Checks designs against Approved Documents and technical standards.
Register interest →Coming SoonBacked by the organisations that set the standard for responsible AI deployment in the UK, including the MHCLG PropTech Innovation Fund.
The technology
One engine. Four products. The full lifecycle
Gpeto's patent-pending architecture combines computer vision, large language models, retrieval-augmented generation and the proprietary PlanCraft engine into a single intelligence layer. It reads planning documents, architectural drawings and site plans, checks them against regulatory frameworks, and adds local context automatically.
The intelligence layer sits below every product, which means each new application, from application creation to validation to S106 delivery to building regulations, inherits the infrastructure already proven in deployment.
Platform architecture
Why vertical AI
The argument for depth
General-purpose AI models are trained on everything. That makes them useful for many things, but trustworthy for very few. With the UK government targeting 1.5 million new homes and MHCLG driving planning reform, the need for trusted, domain-specific AI has never been more urgent.
Domain-specific training data
Models trained on real planning decisions, building regulations, local plan policies, Approved Documents and domain-specific terminology from the built environment.
Validated against real outcomes
Every model is tested against historical planning decisions and known compliance outcomes. We measure accuracy against real-world results, not benchmarks.
Deployed in production
Our first product is live and used by paying customers. The technology is proven in deployment, not just in research.
How it works
The Gpeto Intelligence Core
Computer vision
Reads documents, drawings, site plans and architectural elevations. Extracts structured information from unstructured inputs.
LLM reasoning
Checks extracted information against national and local regulatory frameworks. Identifies compliance issues and policy conflicts.
Geospatial data
Adds local context: conservation areas, flood zones, neighbouring constraints, site-specific policy overlays.
Patent pending.
Responsible by design
Built for trust, not just performance
Explainability
Every assessment includes the reasoning chain: which planning policy or building regulation was checked, what was found, and why the conclusion was reached.
Audit trails
Complete record of inputs, processing steps and outputs. Every decision can be reviewed, questioned and verified after the fact.
Human in the loop
Gpeto assists decisions. It does not make them. Every output is designed to support professional judgement, not replace it.
About
We build AI for the decisions that shape the built environment
Gpeto AI was founded in 2023 with a conviction that the most impactful applications of artificial intelligence are not the most general, but the most specific. And that the built environment is where that specificity matters most.
Our founder, Harold Cabrera, experienced the UK planning system first-hand and saw a problem hiding in plain sight: intelligent, capable professionals spending weeks on manual review that should take minutes. Planning approvals, building regulations, design compliance. The processes that shape every structure around us were still running on paper and instinct.
That experience became the blueprint for Gpeto AI. Rather than building another horizontal AI platform, we chose to go deep into the built environment: learning the regulations, the workflows and the edge cases until we could build something that works in deployment, not just in a demo.
Our mission
To build the intelligence layer for the built environment, so that planning approvals, building regulations and design decisions are made faster, more accurately and with far less friction. Starting with the UK market we know best, then taking our B2C products to homeowners and self-builders around the world.
News & Recognition
Grants, programmes and milestones
Gpeto has built strong institutional backing across the UK's innovation ecosystem. Selected for the MHCLG PropTech Innovation Fund to build S106 Intelligence, piloting with local councils in consortium with PUBLIC. Multiple rounds of Innovate UK Smart Grant funding to develop its core intelligence platform, alongside support for advanced model training and computer vision. Accepted into several prestigious programmes, including the BridgeAI Programme run by Digital Catapult, the Alan Turing Institute, and the GLA Open London Accelerator.
Contact
Work with us
Whether you are an investor, enterprise partner, government buyer, or interested in joining the team.
Join the team.
We are building AI products for the built environment: planning validation, design tools and building regulations compliance. We are looking for engineers, domain experts and commercial talent who want to work on problems that matter.
info@gpeto.ai →
"We are building the intelligence layer for the built environment, so that planning, regulations and design decisions are made faster, more accurately and with far less friction."
Talk to us