- Fiserv, a Fortune 500 fintech firm, has adopted AI-powered surveys to improve customer experiences.
- It can now solve for inefficiencies, its VP of customer insights says.
- This article is part of "CXO AI Playbook" — straight talk from business leaders on how they're testing and using AI.
Fiserv could be considered one of the original fintech companies.
Founded in 1984, the Fortune 500 firm, which has roughly 40,000 employees worldwide, plays a critical role in making sure that payments go through and that people and businesses can access their money.
Despite its vast reach, Fiserv often operates unnoticed — and that's intentional, said Ilana Boyum, the company's vice president of voice of the customer and customer insights.
"Whether you're using mobile banking or working with a local bank as a small business, you're likely interacting with Fiserv," she said. "But when we do our job well, we stay behind the scenes, making everything happen."
While Fiserv stays in the background, customers still need a way to alert the company when something isn't working. Boyum describes her role as "the liaison and translator," helping Fiserv understand its clients' concerns. "Clients know their business, not ours — and that's fine. But they need to tell us what they're experiencing," she said. "They might sense something is off but not know if it's by design or if something is actually wrong."
Fiserv has long relied on customer surveys to identify issues, but vague, one-word responses often obscured the real problem. Boyum recently integrated conversational AI into the process, turning surveys into a two-way conversation. "Now we're cracking the code on what needs fixing and what's working as intended," she said.
Business Insider spoke with Boyum about how Fiserv is using artificial intelligence to gain deeper customer insights, identify inefficiencies, and enhance the customer experience.
The following has been edited for length and clarity.
Before implementing AI, what were the biggest limitations of your traditional customer feedback process?
Historically, each client received three surveys in their first year and then one additional annual survey. If they contacted customer service, they would also receive a survey. Typically, we would keep surveys to a maximum of 10 questions and limited open-ended questions to three.
We always started surveys with a scaled response, followed by a "why." Digging deeper required multiple follow-ups, and responses were often too brief or unclear, making it difficult to uncover the root issues.
Survey fatigue is a challenge for many companies. What do you think contributes most to it?
The biggest driver of survey fatigue is when customers feel like they're giving their time and intellectual energy only to see nothing come of it. That's not always our fault, but it is our problem. We're careful not to ask questions we already know the answers to or have no intention of acting on.
How did AI change that?
When we first started with AI, it provided canned follow-ups for vague answers. It helped, but some questions felt awkward or out of place.
Now we use Qualtrics conversational intelligence, which has a natural flow. It generates adaptive, intuitive follow-up questions so we can drill down and get more robust answers. It's also allowed us to shorten our surveys because we're able to get much greater insight without having to ask a bunch of questions that may or may not apply.
What have you noticed about response rates and the depth of feedback you're getting?
Our response rate hasn't increased, but the quality and actionability of responses have improved. About 40% of customers who completed the surveys with AI-generated questions gave us more detailed feedback.
Before AI, the first answer our customers would give us was often rather vague. They would respond with feedback that was pretty general, like "customer service" or "pricing." Using generative AI to dig deeper and ask conversational follow-up questions has allowed our clients to more accurately articulate what they really mean.
Sometimes the feedback they give is the result of us not properly setting the expectations for how to interact with us. Other times, it is because something is actually "broken." AI has allowed us to tease out this difference.
Can you give an example?
As a financial services company, we work closely with law enforcement, the IRS, and regulatory agencies. So, for example, when funds are held due to an investigation, we can't disclose the reasons for safety and legal purposes. But that can be frustrating for clients. They ask: "Why didn't I get my money? And why can't the agent tell me?"
This was an area we hadn't handled well before, but we've since improved our communication based on what we've learned from our surveys. We let clients know there's a team that will follow up, along with a clear timeline for when they can expect updates.
What measurable impact has AI had on your business since its adoption?
We've seen significant improvements in our customer experience metrics. Each point increase in our net promoter score directly translates to a revenue bump for us. We've closely correlated these, and since implementing conversational AI, we've seen a solid 10-point improvement in NPS scores across our surveys that measure the purchase experience, the activation experience, and the first-90-day experience for our customers. For us, that's translated into better client retention and millions of dollars in revenue.
