Dave Excell

Featurespace developed machine learning technology for real-time fraud and financial crime detection, building its ARIC Risk Hub platform on adaptive behavioural analytics that models the normal behaviour of individual customers and flags anomalies in real time. The Cambridge-originated company was a spin-out from Cambridge University engineering research and applied Bayesian machine learning methods to detect fraud patterns that rule-based systems miss.\n\nThe company raised approximately $108 million including a $108 million Series D round before being acquired by Visa in 2024. Prior to acquisition, Featurespace counted HSBC, Contis, Worldpay, and multiple tier-one banks among its clients, with the ARIC platform protecting hundreds of billions of dollars in transaction volume annually. The acquisition gave Visa proprietary fraud detection AI to deploy across its global payment network and differentiate its data services business.\n\nFeaturespace competed against established fraud management vendors including FICO, SAS, and Fiserv, as well as AI-native challengers including DataVisor and Sardine. Its differentiation came from the ARIC adaptive analytics approach, which modelled individual behaviour rather than relying on population-level fraud rules, achieving lower false positive rates than competitors on several published benchmarks. Integration into the Visa network represents a significant distribution expansion that would not have been achievable as an independent vendor.