Build. Test. Prove. Deploy
Applied R&D that turns uncertain ideas into evidence before material capital is committed.
Labs exists because a demonstration is not the same thing as a dependable operating capability. New AI, forecasting, simulation, robotics and automation concepts need to survive contact with real economics, real data, real users and real controls before a business should scale them.
Labs converts operating questions into testable hypotheses, proves the critical unknown, measures the result and stops weak ideas early.
127.5% Growth in private AI investment in 2025 according to Stanford HAI Stanford HAI, Apr 2026
80% vs 37% In McKinsey's 2026 survey, 80% of respondents said AI improved their individual productivity, while 37% said AI contributed positively to enterprise EBIT. The gap reinforces the need to prove operating value before scaling McKinsey &, Company, Aug 2026
The problem
Companies are surrounded by new technology but often lack a disciplined way to decide what deserves investment. Pilots can become demonstrations that consume time without changing the business, while promising concepts are sometimes scaled before reliability, control or unit economics have been proved.
Our view
The purpose of Labs is to reduce the cost of being wrong. A hypothesis should be made explicit, the critical uncertainty should be tested, and the decision to continue should be based on evidence rather than enthusiasm.
An experiment earns the right to scale. Scale is never assumed.
What Labs does
Labs can explore AI agents, forecasting, predictive systems, simulation, digital twins, computer vision, robotics, industrial data and new forms of human-machine decision support. Markets and forecasting research also sit here when they are being used as research methods rather than offered as a separate we product.
The theme matters less than the discipline. Does it work? Does it create value? Can it be controlled? Can it operate reliably? Can it scale economically?
The lab process
Start with an operating constraint or decision question. Form a testable hypothesis. Simulate where possible. Build only enough technology to test the highest-risk assumption. Test performance, safety, security, usability and economics. Decide whether to stop, redesign or continue. Engineer for production only after the evidence threshold is met. Transfer ownership into operations with monitoring and documentation.
What you get
The customer gets a lower-cost way to test uncertain technology before making a larger commitment. A successful experiment can move into production with clearer evidence. An unsuccessful one can be stopped before it consumes material capital or creates operational risk.
Why North Axiom
Labs is grounded in a long-running habit of testing forecasts, signals and emerging technology against real decisions rather than treating research as an end in itself. That research discipline is combined with practical operating knowledge across regulated, industrial and owner-managed environments.
The purpose is not to reproduce a trading business inside North Axiom, but to bring disciplined experimentation, scenario thinking and evidence into operating problems where uncertainty needs to be reduced before capital is committed.
Where Labs sits in the 8-stage model
Labs supports Digitise, Automate, Predict, Optimise and Scale when a portfolio company or client needs a method that is not yet proven. It is an R&D and validation capability behind the operating model, not a separate investment thesis.
Illustrative example
Illustrative example, not a North Axiom result. A manufacturer considering predictive maintenance may not need an enterprise platform first. Labs can test whether a small set of machine signals predicts a defined failure mode with enough reliability and economic value to justify broader deployment. If the evidence is weak, the concept stops. If the evidence supports it, it moves into the normal Industrial and Predict stages.