Socure has disclosed performance figures for DocV Omega, its document verification model combining biometric and document evidence with device, network and identity context. The company says the model improves fraud detection and response times compared with its Predictive DocV system, which incorporated technology from Socure’s $70 million acquisition of Berbix in 2023.
In its comparison, Socure says DocV Omega captured 98 percent of fraud within the riskiest five percent of transactions, with a false-reject rate below one percent. It also reports 50 percent more fraud capture involving synthetic and counterfeit documents, and 43 percent better detection of injection attacks and deepfakes.
The company says latency fell by 50 percent, with the 95th-percentile response time as low as 250 milliseconds. A 95th-percentile measurement describes the response time within which 95 percent of requests are completed. The results are Socure’s comparison with its earlier system.
DocV Omega connects more than 2,400 predictive variables across 10 categories of information. Its model evaluates how document forensics, biometric matching, device activity, phone risk, network information and other evidence reinforce or contradict one another, rather than relying on each check in isolation.
The output is a calibrated fraud score between zero and one, accompanied by more than 100 reason codes and access to underlying data. Customers can set thresholds and route transactions to approval, manual review, additional verification or rejection according to their policies.
Socure’s RiskOS platform can carry information from earlier steps into the document check. Account recovery after a failed passcode challenge, initial onboarding and an in-person verification can present different risk conditions. DocV Omega is designed to evaluate that context within the model.
Socure says its identity graph has developed over more than 12 years, linking names, addresses, devices, IP and ISP data, documents, biometrics and capture behavior. Signals such as velocity, consortium data, phone risk and IP address, which previously ran alongside document verification without shaping its outcome, can now contribute directly to the composite score.
The company says the relevant context also differs by use case and industry. An in-person verification has a different threat profile from a remote check, where injection attacks are possible, and age assurance presents different risks from account takeover or new-account onboarding. Fraud patterns and risk tolerance also vary across gaming, fintech and banking.
Sources: Socure, Socure Documentation
–
By the ID Tech Editorial Team
