The rapid spread of agentic AI is forcing organizations to rethink identity verification. No longer a one-time checkbox at onboarding, identity management must now run through the entire customer lifecycle, according to security solutions provider Entrust. The firm's 2026 fraud report points to a sharp rise in deepfake attacks that the human eye cannot catch.
The double challenge: humans and AI agents
Online verification systems now encounter two types of users - humans and autonomous AI agents. Consumers deploy AI agents to handle shopping, travel bookings, restaurant orders, and banking transactions. This twin influx puts pressure on organizations that need to deliver a fast, frictionless experience while confirming who, or what, is on the other side of a transaction.
"The real challenge for organizations is how to strengthen compliance and fraud protection while still giving customers a fast and smooth onboarding experience," said Harvinder Singh, Regional VP of Identity Security Solutions at Entrust. "But the way AI is scaling now, the verification of identities is no longer a checkbox activity anymore."
Moving beyond day-one checks
Most institutions have treated identity proofing as a front-door event. Singh argues that view is obsolete. Account takeovers, phishing of one-time passwords, and convincing synthetic fraud require verification to persist across the relationship. "My argument is that we need to look at identity from a life cycle perspective," he said. "I think it is now about building a security identity framework within the system so that I can be trusted at all times, and also so I can trust the system as well."
This continuous approach means weaving fraud signals, biometrics, and device intelligence into every interaction - not just the initial login. With AI agents performing transactions on behalf of consumers, the verification layer must know whom it is serving at every step, not just at account creation.
Inside the fraud lab
Entrust runs an in-house Fraud Lab that replicates attacker techniques to train detection models. "We manufacture frauds from a deep fake biometric perspective and try to stay ahead of the game," Singh said. The lab works on IP address tracking, device binding, and biometric injection attacks. By generating synthetic frauds, the company builds the training data its models need to spot threats earlier in the lifecycle.
Quantum computing will raise the stakes further. Singh believes post-quantum capabilities could produce "absolutely truthful documents and truthful faces which are not human," making the need for AI-driven defence a permanent operational requirement.
Why this matters for management
For leaders overseeing digital operations, the line between fraud prevention and customer experience has vanished. Treating identity as a lifecycle function means reallocating resources from point-in-time checks to ongoing authentication systems. Managers who integrate compliance, fraud protection, and onboarding into a single framework will be better positioned to handle the rise of agentic transactions - and the deepfakes that follow. AI for Management now includes the readiness to deploy and govern these continuous identity controls, not just evaluate them annually.
Your membership also unlocks:
