Know your customer (KYC) and know your business (KYB) remain the first formal barrier between legitimate commerce and organized fraud. Their durability, however, depends on how organizations put them into practice.
Zac Cohen, chief product officer at Trulioo, said too many organizations still rely on episodic controls. “It’s still treating it as a one-time ‘check the box,'” he told PYMNTS as part of a mini-series on blind spots in fraud defenses.
In practice, onboarding, transaction monitoring and account management often sit with different teams. Cohen noted that “different teams will have different responsibilities along the chain of an onboarding cycle or of a risk defense system.” When those functions are not coordinated, controls may look complete within each silo but fall short across the full lifecycle.
The alternative, he said, is continuous oversight. He described an “always on defense” posture where “bad people get friction and good people have a better experience.” Well-aligned controls, in his view, do not hurt usability. They concentrate scrutiny on actors who present elevated risk.
How Signal Depth and Behavioral Patterns Reveal Fraud at Onboarding
Cohen emphasized that not all signals carry equal weight. In both KYB and KYC contexts, he said, “the depth and consistency of the digital footprint” are particularly telling. Authentic identities tend to display layered trails, tenure and behavioral coherence. Fabricated profiles, by contrast, are often thin, recently created or inconsistent across data points.
Businesses add complexity. Newly incorporated firms may lack historical depth. That requires broader data inventories and statistical modeling to assess typical patterns. With the right tools, Cohen said, organizations can score “for the maturity of the business or the identity, as opposed to signal red flags.”
He also drew a distinction between unusual behavior and synthetic conduct. Risky but real users may transact outside expected norms, yet “their behavior still follows like human patterns over time,” he said. Synthetic identities often reveal automation signals, abnormal velocity or mismatches between identity layers and device data.
These insights matter beyond onboarding. Account updates, ownership changes and transaction flows all generate signals. Without integration, those signals remain disconnected rather than forming a unified risk picture.
Why Handoff Gaps Create Structural Vulnerabilities in Fraud Defense
Cohen identified inconsistent risk models as a primary weakness for some firms. “It really comes down to an inconsistent risk model,” he said. Onboarding, profile updates and transactions may run on separate systems, each with different thresholds and rules. That fragmentation creates blind spots when data are not reconciled across platforms.
He also pointed to a drop in scrutiny after approval. Organizations often apply heavier controls at onboarding, then revert to lighter monitoring. The result is not just regulatory exposure. It is structural vulnerability. Fraudsters study those gaps and adjust their tactics accordingly.
How to Balance Fraud Prevention Without Creating a False Positive Machine
Balancing friction and exposure requires more than incremental tuning. Cohen advises executives to view fraud prevention as a lifecycle rather than a single checkpoint. When services work together, the trade-off between conversion and control becomes easier to manage.
Monitoring, screening and surveillance should function as “a network and event driven machine, not binary and batch,” he told PYMNTS. Meaningful triggers include ownership updates, adverse media spikes or newly connected high-risk nodes. Isolated alerts, however, without behavioral or relational context, can overwhelm investigative teams.
By layering watchlist hits with contextual data, organizations can “prioritize true exposure and suppress the low signal alerts,” Cohen said. That approach addresses the most persistent executive complaint: a false positive machine that consumes resources without improving outcomes.
How to Scale KYC and KYB Controls Across Multiple Jurisdictions
Expansion introduces further strain. “Everything changes when you scale,” Cohen said. Risk ratings, cultural norms, customer expectations and data availability all vary across North America, Europe, APAC and other regions.
He advised combining local inputs with universal frameworks so that customization occurs within shared workflows. He also cautioned against relying solely on resold capabilities in new markets. Technology must be built and maintained across jurisdictions. Without that foundation, processes that worked in one region can break in another.
Why AI and Agents Must Run Alongside Existing Fraud Controls, Not Replace Them
Cohen closed with a broader directive for risk and compliance leaders. Emerging technologies, including agentic and AI-driven systems, should run alongside established controls rather than replace them.
“I think the biggest takeaway I’m seeing today is agents and AI technology are not going to replace certain motions,” he said. “They need to run in parallel to enhance our positions and enhance how we manage risk and how we capture fraud and prevent it from entering our ecosystem.”
