Artificial intelligence may be entering an era where the technology itself is less important than how organizations redesign work around it.
This was the central theme of the next edition of PYMNTS Payments, featuring Karen Stroup, Chief Digital Officer of WEX. Rather than treating AI as another feature added to existing software, Stroup argued that companies should rethink the underlying processes and ask themselves what would happen if they started all over again.
Her perspective also reframes the familiar debate about AI as either aspirin that solves immediate problems or vitamins that bring about incremental improvements. While she acknowledged the usefulness of the comparison, she suggested that it captured a fleeting moment in time.
“I seriously look at this as a comprehensive treatment plan, not just aspirin and vitamins,” Stroup said.
A perspective beyond individual functions
Many organizations begin their artificial intelligence projects by looking for a single problem to automate. Stroup believes this approach often limits potential benefits.
“When you think about the true potential of AI, it’s not a matter of urgency at the time of purchase,” she told PYMNTS. “It’s important to rethink the way we work and create solutions that address it all.”
That philosophy influenced the way her team approached product development. Before implementing artificial intelligence, first map your existing customer journey, identify unnecessary steps, and define what your ideal experience looks like.
“We encourage teams to create service design blueprints,” Stroup says. “Map your experiences, talk about what you’re doing today, and dream big.”
The goal is not simply to speed up an existing process, but to determine whether the process itself should exist in its current form.
Measure results instead of tokens
As companies invest heavily in generative AI, many executives are monitoring adoption through rapid counts and token consumption. Stroup warned that such measurements can be misleading.
“How many tokens do you spend? Well, you could spend a lot of tokens and build a lot of things that don’t matter,” Stroup said.
Instead, organizations need to evaluate speed of decision-making, product innovation, customer outcomes and operational improvement, she said. These metrics provide a clearer picture of whether AI is creating business value rather than just generating activity.
The same philosophy applies to your products. To use a feature, users must remember that the feature exists. Functionality built into daily workflows eventually becomes part of the business’s own operating system.
Built around customer needs
That mindset shapes the services WEX is developing for its customers.
In the health and benefits space, Stroup described a future where employees are no longer scrambling to find receipts weeks after paying with flexible spending or health savings accounts. AI can verify information at the point of care, automatically store documents, and eliminate much of the subsequent substantiation process.
The goal is to remove work from the customer rather than imposing new administrative obligations.
Your fuel card becomes an intelligence platform
The same principle manifests itself in fleet management.
Fuel cards generate large amounts of transaction data and can reveal unusual behavior, from out-of-network purchases to repetitive fueling patterns worth revisiting. AI can quickly identify these anomalies and alert administrators before the problem becomes serious.
Stroup said the system could work beyond issuing warnings by creating communications that help managers coach drivers on appropriate behavior. AI doesn’t just present information; it supports next operational steps.
This combination of insight and action shows how payments products are evolving from transaction tools alone to decision support systems.
The value of dreaming bigger
Stroup also warned companies not to wait to experiment with artificial intelligence until they think it’s fully established.
“The arrival of AI is inevitable,” she said. Organizations that start learning develop experience with data, workflow, and governance, even when individual experiments fall short of expectations.
For her, the greater risk lies in maintaining yesterday’s processes while competitors redesign tomorrow’s.
Read Karen Stroup’s full interview to learn more about:
- Why companies should expect the experiments themselves to create value for the organization, even if individual AI projects fall short of their original goals.
- How embedded intelligence can transform year-end benefit planning by helping employees make better funding and enrollment decisions before problems arise.
- Why Stroup believes AI will move managers from performing mundane business tasks to considering recommendations and making higher-level strategic decisions.
