Page 35 - gs260802

Insights and Expertise

gateway can route transactions across acquiring paths
with higher probabilities of approval. Over thousands of
Smart gateways require smart questions daily transactions, even small improvements in accep-
tance rates accumulate into meaningful revenue recovery.
As AI becomes embedded in payment gateways, evaluating
the technology may require merchants and their payments The competitive pressure to adopt these capabilities is be-
partners to look beyond a provider's list of AI-powered ing amplified by the behavior of the networks themselves.
features. The more consequential question is how reliably Visa and Mastercard both launched agentic AI payment
those features perform in a live payments environment. frameworks in the latter part of 2025, with Visa's Trusted
• What data informs the system? An AI tool is only Agent Protocol attracting partners including Stripe, Shop-
as useful as the information available to it. Transac- ify, Worldpay, Microsoft, OpenAI and Anthropic.
tion history, issuer responses, fraud signals, routing
outcomes and other data can give a gateway a richer PayPal and Mastercard subsequently deepened their part-
basis for identifying patterns. Merchants should un- nership to advance agentic commerce, targeting hundreds
derstand which data the system uses, how current it is of millions of consumers and tens of millions of merchants.
and whether insights are based on their own activity,
broader network intelligence or both. When the largest networks in the world are restructuring
their protocols around AI-native interactions, a gateway
• Can it explain its recommendations? Natural-lan- without AI capability is not just behind the curve; it's op-
guage explanations are one of the attractions of LLM erating in a framework that was not designed for it.
technology, but a convincing explanation isn't neces-
sarily a correct one. When an AI system recommends How merchants benefit in practice
changing a route, retrying a transaction or responding
to an emerging pattern, users should be able to under- The value of LLM-powered gateway technology flows to
stand the evidence behind that recommendation. merchants across three distinct dimensions. The first is op-
• Where does human judgment enter? Automation can erational clarity. Instead of receiving opaque decline codes
accelerate routine decisions, but not every payment that require a technical specialist to interpret, merchant
problem should be resolved without human review. teams receive natural language explanations that can in-
Providers should be able to explain which actions form immediate action, whether that means requesting a
the system can take autonomously, which require ap- card retry, updating payment credentials or identifying a
proval, and how unusual or high-risk situations are systemic routing issue.
escalated.
The second dimension is strategic intelligence. An LLM
• How is sensitive information protected? Payment
systems handle data that demands stringent security layer trained on a merchant's transaction history can sur-
and privacy controls. Adding AI introduces another face trends in real time, flagging performance deteriora-
layer to the technology stack, making questions about tion across specific card types or geographies before it be-
data access, retention, model training and third-party comes a material problem. This kind of proactive insight
exposure increasingly important. was previously the preserve of merchants large enough
to employ dedicated payments analysts. AI-powered gate-
• How is performance measured? Claims about intelli- ways democratize access to it.
gent routing or fraud detection matter most when they
translate into measurable results. Merchants and their The third dimension is resilience. As fraud tactics evolve
payment partners can look for improvements in au- and regulatory requirements grow more complex, an AI
thorization rates, false declines, fraud losses, process- infrastructure that learns from new signals and adapts
ing costs and the time employees spend diagnosing continuously outperforms one locked to static rules. The
payment problems. merchants best positioned in the payments landscape of
As AI becomes embedded in payment gateways, eval- 2026 and beyond will not simply be those with the highest
uating the technology may require merchants and transaction volumes. They will be those whose payment
their payments partners to look beyond a provider's infrastructure is intelligent enough to improve itself.
list of AI-powered features. The more consequential
question is how reliably those features perform in a The gateway is no longer just a conduit. It is becoming the
live payments environment. most consequential piece of technology in a merchant's
commercial stack, and the organizations that recognize
this shift early will hold an enduring structural advantage
over those that do not.
direct financial benefit. Every declined transaction that
could have been approved represents lost revenue.
Andrii Shevchuk is CTO and partner at CONCRYT, https://concryt.
By using AI to analyze issuer behavior, card type, geogra- io. Contact him via LinkedIn at linkedin.com/in/andrii-shev-
phy and transaction context simultaneously, an intelligent chuk-53224645.
35