Every consequential decision should be simulated before it’s made.
The world got faster. The way we make decisions got quicker, not necessarily better. AI didn’t fix bad decisions. It scaled them. The same partial inputs, processed at machine speed, by systems that are useful but often missing the context and judgement a real decision needs.
Timelines compress, ownership diffuses, and organisations end up with a dangerous illusion: more speed and more confidence, without a matching increase in accuracy. We started Cambium because we believe the winners of the next decade won’t be the ones who move fastest. They’ll be the ones who get the decision right, faster.
Decision simulation, built on a population that actually exists.
We’re not another model that answers faster. We reconstruct the US population, individual by individual, and the places those individuals move through, so businesses, investors and policymakers can pressure-test a decision before they commit capital, products, or policy to it.
- Test the price rise before you launch it.
- Test the market entry before you fund it.
- Test the policy before it reaches the public.
Not a single prediction or a single opinion, but a distribution of real-world outcomes you can query, challenge, and learn from.
A synthetic population engine, built entirely from public and licensed data.
No personal data. GDPR, CCPA and HIPAA compatible by design. Around 15,000 columns of public, administrative and lifestyle data are fused and re-weighted down to US Census block-group level, then calibrated against real-world outcomes.
On top of that engine sits a decision layer, built vertical by vertical, starting with product management and extending into government & policy, financial services, digital marketing, and political intelligence. Every simulation run makes the next one sharper: a data flywheel that gets more precise the more it’s used, built to empower expert judgement rather than replace it.
Built by people who have shipped decisions at scale.
Michael Birdsall
Michael has spent his career turning data into decisions that hold up under pressure. Chief Data Scientist through MarketWise’s £2.35B IPO, Chief Data Officer at Tes Global through its £850M PE exit, and founder & CEO of TwoSigmas to a 44x acquisition exit.
Cambridge MBA · Wharton PE · MIT ML · US Navy Nuclear Engineering
Adelle Wood
Adelle has built go-to-market from the ground up more than once: part of the team that launched Peloton’s UK Studio to £1m+ in year-one retail sales, driving AI-led customer expansion at MarketWise, and marketing at TwoSigmas.
Stanford GSB, Digital Transformation · Google Squared, Digital Marketing
Josh Gillott
Josh builds the systems that make the impossible tractable. As a Data Scientist at Alphapeak he built the RAG chatbot that opened strategic partnership talks, and arbitrage models that lifted yield 66%. Previously a Data Analyst at Tes Global.
MMath in Applied Maths, Statistics & Quantum Mechanics, University of Sheffield
Gavin Ferris
CEO, Common AI. Cambridge PhD in AI, with hands-on experience scaling AI ventures from research to market.
Peter Hiscocks
Serial entrepreneur (Pod Point, Cambridge Enterprise). Built the entrepreneurship programme at Cambridge Judge Business School; investor and deep-tech advisor in the Cambridge venture ecosystem.
Phil Mansell
Former CEO, Jagex Games. Operating experience scaling a global product business; advises on leadership, org design and operational discipline.
George Prest
Part of the team behind The Brandtech Group’s growth to a $4.1bn valuation. 25+ years as writer and creative leader across narrative, campaigns and brand; former global creative lead at Lowe and R/GA.
Get started today
Ready to reason about real people?
Stop asking AI to imagine your audience. Start with populations reconstructed from verified public data.