As the digital landscape shifts under the weight of AI-driven interactions, the traditional methods of establishing trust via humans has changed. The identity industry was built for human users, but that assumption is now broken.
Trust models designed to verify humans at a single moment in time are no longer sufficient. AI agents, deepfakes, and automated fraud now operate continuously, at machine speed, exposing the limits of one-time identity checks.
Microblink, a global leader in identity intelligence, closed a record-breaking 2025, processing 2.9 billion identities and achieving 3× growth in verification volume. Leading this charge is CEO Hartley Thompson, who argues the industry must evolve beyond one-time “Know Your Customer” (KYC) verification toward a new paradigm: “Know Your Actor” (KYA), a model built on continuous behavioral assessment across the full digital journey.
We sat down with Thompson to discuss this strategic pivot, Microblink’s Identity Intelligence OS, and how businesses can maintain control in a world where AI-driven actors are becoming the norm.
Q: Hartley, Microblink is officially ushering in what you call the “Know Your Actor” era. How does this concept fundamentally differ from the traditional “Know Your Customer” model. KYC is identity at rest, verifying a person, vs. KYA, which is identity in motion, evaluating authority, behavior and context continuously. Why is this shift necessary right now?
Hartley Thompson:
For years, digital trust has operated on a flawed assumption: verify a person once and you can trust everything that follows. That model worked when every interaction was human-initiated and relatively contained. It doesn’t work anymore.
Today, humans, AI agents, and automated systems all operate inside the same digital environments, often acting on a user’s behalf without direct human involvement. Most identity frameworks were not designed for that reality. They were built for a point-in-time decision, not continuous evaluation.
That’s why the industry has to move beyond Know Your Customer. The real question is no longer just who a person is. It’s what or who is acting for them in a given moment, whether that authority is legitimate, and whether it should still be trusted as behavior evolves. That’s what we mean by Know Your Actor. Trust has to become continuous, contextual, and adaptive, or it breaks under the scale of automation.
That’s no longer the environment organizations operate in.
Today humans, bots, and AI-driven agents all interact inside the same digital flows. An assistant can search, book, authorize, modify, and transact on behalf of a person, sometimes without the human directly participating in that moment. So the question changes. It’s not just who is the person — it’s what is acting for them, whether it’s authorized, and whether its behavior still matches that authority over time.
Know Your Actor reflects that shift. Instead of a one-time verification, trust becomes a continuous evaluation across the lifecycle of an interaction. Identity risk now evolves and fraud doesn’t occur at a single checkpoint, so organizations have to evaluate behavior and authorization, not just identity at onboarding.
We’re seeing this urgency driven by agentic AI and automation. The core decision businesses now face is no longer “Was this person verified?” but “Should this interaction be trusted right now?”
Q: You’ve reported impressive metrics for 2025, including processing nearly 3 billion global identities and seeing a 50% increase in platform adoption. What specific market forces or client needs are driving this surge in demand?
Hartley Thompson:
What’s really driving demand is a change in how fraud works and how digital experiences work.
Fraud is no longer limited to onboarding, and identity risk doesn’t stand still. Generative AI has made it dramatically easier to create convincing fake documents and synthetic identities, and those tools are becoming widely accessible.
At the same time, companies are digitizing more of the customer journey and automating workflows across marketplaces, financial services, and hospitality. That increases scale, but it also increases exposure.
Businesses now have two pressures at once. They need lower friction and faster user experiences, but they also need stronger protection. Static checks were built for a slower, human-only environment. As automation grows, those checks leave gaps between systems, and attackers exploit the gaps rather than the controls themselves.
So customers are looking for a unified way to evaluate trust across onboarding, authentication, and transactions, not just a compliance step. The growth we’re seeing reflects a broader shift from identity verification to ongoing trust decisioning across the entire journey.
Q: One of the key challenges mentioned in your recent announcement is the fragmentation of identity tools. How does Microblink’s “Identity Intelligence OS” address this issue for enterprise leaders?
Hartley Thompson:
Most enterprises didn’t design their identity stack all at once. They layered it over time through document verification, biometrics, device intelligence, fraud tools, payment checks, etc., often from different vendors. Each system makes its own trust decision, which creates silos and inconsistent outcomes. Modern fraud attacks those seams between systems rather than any single control.
The Identity Intelligence OS is meant to coordinate those signals into one decisioning framework. It unifies document, biometric, fraud, and payment signals under centralized logic so trust can be evaluated continuously instead of independently. Instead of separate pass-fail checks, organizations can evaluate behavior holistically and adapt decisions as risk changes over time. The goal is explainable and configurable outcomes rather than fragmented approvals.
For enterprise leaders, it becomes a single trust layer operating across onboarding, authentication, and transactions without forcing them to rebuild their infrastructure.
Q: With AI-generated fraud rising by 23% last year, static identity checks are becoming obsolete. How is Microblink using its “Fraud Lab” and continuous risk signals to stay ahead of threats like synthetic IDs and deepfakes?
Hartley Thompson:
The fraud problem changed quickly. Producing a realistic fake ID used to require specialized expertise. Now anyone with an internet connection can generate a convincing deepfake document using widely available tools.
We created the Fraud Lab to be proactive rather than reactive. The team generates synthetic identity documents, recreates deepfake patterns, and models real attacker behavior so our systems are trained on emerging threats before they reach customers.
The lab produces more than 100,000 synthetic IDs per month and tests thousands of attack variations, including face-swap IDs and fabricated documents that visually appear legitimate.
Instead of waiting for fraud to appear in production logs, we try to understand how attackers operate and strengthen detection in advance. Then continuous signals like behavior, biometrics, and transaction context allow organizations to respond when risk changes, not just at account creation.
One example we’ve seen is a fintech provider serving credit unions that is now detecting more than 1,300 fraudulent identity documents every week using capabilities developed through the Fraud Lab. These were cases that previously could have passed onboarding because the documents appeared legitimate. What it shows is how quickly fraud techniques have improved and why reactive controls are no longer enough.
Q: Looking ahead, you’ve noted that we are entering a world where humans and AI agents will interact in the same digital spaces. How is Microblink preparing to facilitate trust in this mixed environment?
Hartley Thompson:
We’re moving toward a digital world where humans and software agents operate together. Agents may initiate transactions, modify services, or interact across platforms on a user’s behalf.
The challenge becomes verifying authority and maintaining accountability. Systems have to confirm the interaction is connected to a verified individual, that the agent is authorized to act for them, and that its behavior stays within permitted boundaries.
Our approach is continuous identity control. Instead of identity being a checkpoint, it becomes an ongoing trust layer across platforms, devices, and automated systems.
The goal is to let automation scale while keeping accountability and user protection intact as interactions become increasingly autonomous.
The companies that win in an AI-driven economy won’t be the ones that verify the most identities. They’ll be the ones that continuously understand who or what is acting in their systems — and whether it should be trusted.
Microblink’s vision for 2026 and beyond is clear: trust requires continuous, intelligent control. By unifying document, biometric, and data verification into a single platform, Thompson and his team are equipping organizations to distinguish between genuine users and AI-driven threats with unprecedented precision. For more information, visit Microblink.com.
