There’s a lot of talk about artificial intelligence, including in the payments community. But experts on a Smarter Faster Payments Remote Connect 2026 panel offered a reassuring assessment: humans aren’t going anywhere anytime soon. 

“I’m not sure we’re ready for agentic (AI) to run by itself without a human in the loop,” said Jagdeep Sahota, Founder of Luminary Advisors LLC. 

“We’re not there where we can remove a person for operational cost at this point. I would hold back on where the model is making a payment decision or independently deciding to move money, just yet,” said Sahota, who had some compelling reasons for not rushing into things.

“Just one bad payroll being held back, or one misdirected payment is going to give you so much negative news out there, that will have everything from your stock impact to your reputational risk of being on the front page of The Wall Street Journal,” Sahota said.

Chetan Cariappa, Head of Payments - US, at Intellect Design Arena, agreed, saying “you use AI to complement people,” and that “it should help improve the decision-making capability of the bank and not replace the actual person there.”

“Trust is—all puns intended—the currency of the bank. You can’t win that back once your reputation is challenged,” said Cariappa. “Wading in rather than diving in is really key in the AI world. And setting up guardrails. Whatever decision-making capabilities you’re going to build into AI, I think putting guardrails around that is a key piece of that.”

Sahota said to “take some baby steps—get into AI, work with it, improve on the operational efficiency without taking people out of the equation right now.” 

So where is AI making a difference for payments today? Exception handling is high on Cariappa’s list.

“Every payments-related organization spends a huge effort in investigating exceptions,” said Cariappa. “A lot of that tends to be following a repeatable pattern. Where AI can play a role there is it’s really good—when done right—in identifying issues. It’s really good in recommending corrective action, in gathering supporting information and data related to that issue.” And once AI has access to all the necessary information, it “can resolve problems in a fraction of the time that it takes a manual process to do.”

Sahota said fraud decisioning is another great use of AI.
“It can gather a lot of data if you have access to real-time data, it can bring it all together and make a much more informed decision than someone clicking through three or four different screens or three or four different tool providers,” said Sahota. 

But what happens to the people in payments? Mark Dixon, Senior Consultant at Nacha Consulting, likes the idea of seizing on someone’s subject matter expertise and bringing them into an analyst role “where we’re really saying, ‘take your experience, take the things you know really well, and make sure we’re driving forward correctly, that the AI is working right.’” 

“And not only that, that we’re taking the data from those models and using it effectively, because everything in business is based on data and effective data usage,” said Dixon. “You need a person to be able to do that side of it—at least today.”