Payments and credit card service organizations have always been measured by their ability to reduce friction. Contact center average handle time, dispute resolution speed, and fraud detection accuracy are all metrics that shape strategies and incentives.
The rise of generative artificial intelligence (gen AI) promises a massive improvement across each of these key performance indicators (KPIs). But while speed and efficiency have long dictated the terms of competition in payments, trust and humanity have always sat at its center.
“As new technologies emerge, whether it’s ML, AI or gen AI, we’ve always approached them thoughtfully,” Gary Kensey, EVP and Unit CIO of Global Servicing & Corporate Technology at American Express, told PYMNTS. “We enhance our customer and colleague experience while staying true to our brand promise.”
That promise, Kensey added, is rooted in trust, security and service.
And as gen AI races into every corner of business, Amex is intent on adding speed and personalization without knocking those pillars out from under its brand.
Because when someone picks up the phone or opens a customer service chat, it’s usually not because things are going well. It’s because something went wrong: an unauthorized charge, a declined payment at a critical moment, or confusion over billing.
Advertisement: Scroll to Continue
In those moments, reassurance matters more than raw speed.
“The North Star is simple: AI should work in service of people, not instead of them,” Kensey said. “We’re not really interested in just the traditional low-hanging fruit of AI to enhance efficiencies. We want to strengthen what we feel is a world-class customer service that defines our company.”
Moving AI Beyond Efficiency
If AI is deployed simply to remove people from the equation, companies may risk eroding trust at the very moment it’s most fragile. That’s why forward-thinking firms are adopting a different playbook, one where technology works behind the scenes to make human service faster, smarter and more empathetic.
Amex’s own view is informed by fresh consumer research the company has been studying. The takeaways are unambiguous: Customers welcome speed and convenience, but they also want to feel understood, and getting to a real person when it matters is crucial.
According to Amex’s research, 70% of surveyed U.S. adults have interacted with some sort of gen AI-powered service tool. Nearly three-quarters say their experience was positive. However, more than half remain concerned about potential lack of human empathy and understanding in a gen AI-powered customer service world.
That tension runs through nearly every brand’s AI strategy right now. On one side sits personalization that feels tailored and timely; on the other, the risk of interactions that feel robotic.
“We want the speed, the efficiency, the automation, but we really need to balance that with trust, empathy and care, the core expectation of our customer base,” Kensey explained. “Adopting gen AI won’t come at the cost of that human connection.”
In practice, that means using AI behind the scenes while keeping skilled people firmly in the loop. Amex is streamlining back-end workflows by using AI to summarize voice interactions, identify the right knowledge article, or propose the next best actions, and reserving humans for the more complex tasks.
“AI can anticipate needs, surface insights, and streamline workflows and processes so our Customer Care Professionals can focus on things like empathy and that human connection,” Kensey said. “If we get it right, the long term is about building deeper, more trusted relationships at scale.”
The cultural shift may be the hard part.
“If the future is to be as successful as we want it to be, the culture has to evolve too,” Kensey said. “You have to build AI confidence in your company’s culture … empowering both the technical teams and the rest of the company to embrace emerging technologies.
“Focus on your own personal digital dexterity and understanding of how to use these tools to make what you do easier and more effective,” he added.
Relationship-Powered, Tech-Enabled
Amex’s company mantra, “relationship-powered, tech-enabled,” isn’t so much a slogan as an operating model that guides investment.
“Customers want digital convenience for simple tasks, but they also want the reassurance of human connection for complex scenarios,” Kensey said, stressing that user choice remains a design principle.
“Whether through chat or phone or digital self-service, we want to make it extremely simple to get to an agent or escalate to a human. That shouldn’t be a burden upon our customer.”
One tangible example is voice summarization in contact centers, while another is the company’s internal knowledge hub. These aren’t moonshots; they’re real workflows that remove friction while improving quality and consistency. The strategic bet is that by freeing people from the retrieval and reconciliation chores, Amex can redeploy human effort to the higher-order skills that customers actually feel: listening, empathy, judgment and reassurance.
And if the first wave of AI in service focused on speed, the next wave aims at relevant delight: knowing not just who you are, but what you need next.
“We’re developing gen AI use cases that will push personalization further by making tailored recommendations in areas our members are most passionate about, whether it’s travel, dining or entertainment,” he said.
“Personalization is built on our closed-loop model and long-term investments in enterprise data capabilities. Gen AI takes it to that next level,” Kensey added, noting Amex’s long history of leveraging customer data safely and responsibly.
“It’s trusted insights from millions of premium card members and from millions of merchants. This allows us to personalize not only at scale, but with a level of accuracy that, without that closed loop, would be really, really hard to achieve.”
If there’s a thread that permeates through Amex’s AI playbook, it is Kensey’s favorite: reassurance. The company wants customers to know that their needs will be met quickly and correctly; that a real human is there if the going gets complicated; that their data is handled responsibly; and that the brand won’t trade empathy for efficiency.
Source: https://www.pymnts.com/