Page 24 - gs260201

Cover Story continued

"First, every automated action must tie back to verified Aggregating diverse capabilities, such as document
human intent," he said. "Today, that means confirming the verification, biometrics, payment card checks and risk
person who authorized the agent. Second, trust requires signals, into a comprehensive framework creates a
continuous behavioral evaluation, not a single credential foundational layer for continuous behavioral evaluation
check. Systems must monitor whether actions remain over time that reinforces trust, Thompson explained.
aligned with permissions over time. Third, trust decisions
must be consistent across channels." He pointed out that the framework functions as a
platform enterprises can leverage to balance security,
Most importantly, Thompson added, agentic and human compliance and growth while delivering an exceptional
trust must be held to the same standard, whether it's a user experience. It also safeguards agentic commerce,
human transacting on a mobile app or an agent buying he added, by keeping agents transparent and auditable
something online. The evolution from human-centric, through clear records of who did what, when, where and
know-your-customer (KYC) compliance to governing under whose authority.
digital actors is a major shift, especially for traditional
fraud models that tend to look for anomalies after a Less friction, more freedom
transaction has been completed. "Every major evolution in commerce brings new risk and
new opportunity," Thompson said. "The organizations
Thompson portrayed that shift in financial services as that thrive will build trust, not just defenses."
moving from confirming trust at a single touchpoint to
continuously reviewing and verifying activity across Sharma concurred, advocating an AI-first approach to
the lifecycle of agent-driven transactions, with scrutiny building sustainable agentic commerce infrastructure.
escalating only as confidence degrades.
"As we move through the AI integration lifecycle, from
"In the old model, you checked someone's ID once at the discovery to prototypes to production, agentic workflows
door," he said. "Now it's about understanding how they and automations become part of the project," Sharma said.
continue to behave throughout their visit." "These agents are present in the prototypes and pilot

Omnichannel identity protection builds that validate an integration before it moves into
production and deployment, replacing isolated tools with
Thompson urged enterprises and teams to invest in what embedded intelligence to remove friction and improve
he called "trust infrastructure," to align with agentic decision-making and relevance."
protocols in payments, technology and commerce. As
Google, Microsoft and other major players build protocols When intelligence and trust are in sync, the payoff is
that allow AI agents to initiate and complete transactions, human freedom, Thompson added. For example, he
he noted, Microblink and other technology platforms mentioned an agent could be shopping for better insurance
are building digital identity frameworks that verify the and switching carriers late at night while a policyholder
human behind the agent. sleeps; other authorized agents could be autonomously
rebooking flights or trading assets, triggering requests
These digital identity frameworks are always on and for biometric authentication only for higher-risk actions.
always connected, Thompson said, and designed to Continuous trust, he noted, allows friction to appear when
reduce friction by monitoring, detecting and preventing needed without slowing workflows or payment flows.
fraud, deepfakes and synthetic identities across multiple
channels and media, which he described as follows: "This gives me more time to spend with my family, build
a business or travel," Thompson said. "The net positive
• Document verification: Extract data in real time is humanity having more freedom instead of scrolling
and in multiple languages using AI-powered non- endlessly to transact."
identity document extraction and classification.
As Sharma and Thompson made clear, when intelligence
• Payment card verification: Perform liveness and trust operate as core components of a dynamic,
checks for card-present transactions, reducing adaptive system, agentic commerce can move from a
chargebacks and ensuring PCI-DSS compliance. technical construct to a material reality. Both agreed

• Orchestration and risk intelligence: Combine autonomy is within reach, but success will depend on
document, device and behavioral signals, using how deliberately intelligence and trust are designed to
real-time analytics to assess pass rates, risk signals work together, not as parallel capabilities but as a single,
and operational performance. integrated foundation.
Dale S. Laszig, content strategy director at The Green Sheet and founder
• Continuous AI innovation: Combine on-device and CEO at DSL Direct, is a payments industry journalist, creator and
processing with cloud-based AI models to consultant. Connect via email at dale@dsldirectllc.com and LinkedIn at
maximize speed, privacy and fraud detection www.linkedin.com/in/dalelaszig.
while supporting continuous learning.
24