Construction Management is the highest circulation construction-based publication serving the UK built environment.
Digital Construction
Startup lowdown: Build tackles the pre-construction phase
Justin Stanton Editor, CM Digital
Image: Vladimir Timofeev | Dreamstime.com
Would you like to significantly speed up the pre-construction phase on major industrial projects? Want to deal with technical due diligence and early design work in days rather than months? If so, Build and its army of AI workers may be the answer.
James Stirrat-Ellis, CEO of Build, gives the lowdown on his startup.
CM Digital: What problem does your technology solve and how does it do it?
James Stirrat-Ellis: Every major building project – a data centre, an energy facility, an industrial campus – goes through the same painful pre-construction phase before a single brick gets laid. Site selection, technical due diligence, feasibility analysis, early design work: it involves dozens of specialists, months of sequential work, and a stack of pdfs that get emailed around and routinely ignored. The process hasn’t changed in decades, and it’s the single biggest reason projects are late and over budget before construction even starts.
This is not a paywall. Registration allows us to enhance your experience across Construction Management and ensure we deliver you quality editorial content.
Registering also means you can manage your own CPDs, comments, newsletter sign-ups and privacy settings.
Build automates that process. We’re a ‘neofirm’: a professional services firm where the majority of the work is done by AI workers that can swarm around complex work over hours, with a layer of domain experts who review and deliver the output.
A developer sends us a brief the same way they’d brief a consultant. We run hundreds of agents simultaneously across the relevant data – planning records, grid capacity, substation queues, site ownership, environmental constraints, design standards – and deliver verified, institutional-grade output in days rather than months. The same work that used to take six weeks now takes hours. Ten times the speed, half the price.
OK, who’s tested or already using your tech and what are the results?
We’ve delivered more than 150 projects across 18 countries. Our clients include Tishman Speyer, the UK government, Fortune 500 companies, hyperscalers and the world’s largest data centre developers. Most of our engagements are under NDAs, but the results are consistent.
Development teams that used to spend six to eight weeks on technical due diligence for a single site are now screening multiple sites in days. One of our clients tripled their monthly quota within weeks of starting. That’s how quickly the value compounds once teams start working differently.
The feedback we hear most is that Build changes how teams show up. They go to site visits with critical information that gives them negotiation leverage. They arrive at investment committees with designs and direction already in hand rather than still gathering data. And they work in parallel across multiple workstreams instead of sequencing dozens of specialists one at a time.
We’ve also mapped power capacity across the US and selectively across Europe, so we can identify sites where grid constraints won’t kill a project before it starts. In the current data centre build-out, that’s as valuable as anything else we do.
Tell us about your senior management team’s background
The founding team sits at the intersection of the two problems we’re solving: how the built world works, and how AI actually gets deployed at scale.
I trained as an architect and worked at Heatherwick Studio on major international infrastructure projects. I then moved into AI, building agent systems before co-founding Build. Ben McClusky, CTO, specialises in multi-agent AI systems and reinforcement learning research.
The wider team combines built-world expertise from Tishman Speyer, Starwood Capital and JP Morgan with AI-native talent from Y Combinator, Harvard and other top research institutions like Imperial College London. That combination is deliberate. You can’t build reliable domain-specific AI without people who genuinely understand the domain, and you can’t scale a services firm at software margins without people who understand the technical layer.
Are you looking for funding?
We closed an $8.5m seed round in June 2026, led by Index Ventures, with Pebblebed – whose founders co-founded OpenAI and Meta AI Research – as part of the round. Our angels include the CFO of OpenAI, the CTO of Blackstone and (former F1 World Champion) Nico Rosberg, among others. That, plus our revenue growth, will get us where we need to go next.
