Now Reading
What Amex, Capital One and TIAA learned by applying AI to contact center transcripts

360 Magazine 
in Print

BUY NOW

What Amex, Capital One and TIAA learned by applying AI to contact center transcripts

AI contact center transcripts

You could say that every Voice of the Customer program is an exercise in panning for gold, but in Luis Angel-Lalanne’s case, the treasure was buried in an unusual place.

The vice-president of customer listening in the Global Servicing Group at American Express says his firm is like many others in that its traditional surveys are linked to an incentive program enjoyed by its front-line employees. That often limits the ability to use that survey data for deeper research purposes. But then, one day, members of his team said they had found cold transcripts stored in Amex’s data lake.

“No one told us they were there. We just went and found them,” Angel-Lalanne recalled during a session at market research firm Forrester’s CX North America event in Nashville last month. “My team started exploring the transcripts and then thought like, ‘Hey, maybe there’s something here that we can use to model customer satisfaction, where we could use the transcript and break that linkage with the survey for the frontline incentive.”

After approximately a year and a half of continuing to build more data into the model and selling the concept internally, Angel-Lalanne said the team successfully rolled out a program to replace the frontline incentive at the end of the first quarter.

According to Forrester senior analyst Colleen Fazio, there are similar transcripts sitting dormant and ignored for months or even years on a shared drive in many other organizations. This may be the perfect time to start dusting them off.

“We are now at a moment in history where we can use AI to start making sense out of what was once unfathomable,” she said. “We can make the contact center better with it, but we can also use it in other places in the organization, such as getting a better Voice of the Consumer (program).”

VoC benefits at a glance

Although no one in the CX North America session called out specific applications or platforms they’re using, they agreed that contact center recordings are a good example of unstructured data that AI can parse for business benefits.

At TIAA, for example, it has meant being able to identify the characteristics and behaviors that that lead to a positive or negative experience on its digital, phone and chat channels. Sheila Catenza, TIAA’s senior director of Voice of the Customer, said these insights are then supplemented with additional research.

“We can package this data and share it with our contact center leadership teams, our training teams and our change management teams,” she said, as well as external business partners. “It really allows us to share with them our customers’ feedback, including their sentiment on our products and their experiences.”

At Capital One, text analytics and targeted closed-loop feedback is helping to understand exactly what might be going on in a particular area of the customer experience, said Anne Louise Mason, its senior product manager, Customer Experience Measurement Product Portfolio.

“We have discovered that the frontline insights are easy to get to the frontline. You can make improvements on how calls go, how training goes and how those conversations proceed,” she said. For other areas, such as product managers responsible for a customer-facing app, “It’s helping to show a broader spread of the data, and to show the insights that come from your specific service.”

Introducing business units to customer feedback

Mason said that Capital One first launched its text analytics strategy in 2018, when the U.S. was on the brink of the federal government shutdown and began hearing from customers who were worried about making late payments.

“These were DC government bureaucrats who were used to following all the orders. They’d never even had a library fine,” she said. “They were terrified of being late and what it would mean. And so we were able to leverage those early signals to create a process that pushed deadlines back and ensured that you weren’t going to face any kind of long term financial consequences for something that was outside of your ability to pay.”

See Also
Michael Lawder ASAPP

Angel-Lalanne said Amex achieved something similar when it used customer listening tools and AI to assess the sentiment of a recently-launched financial relief program, the results of which were passed on to its risk management group.

“You know, this is a group that was not used to (receiving) customer feedback, but they’re the ones obviously dictating the terms of the financial relief plan and the whole program,” he said. “It really opened their eyes to the feedback that we’re collecting on the front lines and in call center. And the reason I know we were successful is that even after we dialed frequency of sending daily feedback to weekly, each week I was getting requests to add another five names to the distribution list.”

Catenza said there is a responsibility for those applying AI to contact center transcripts to not only share this data with business partners, but also to make sure that you contextualize and use storytelling to bring the information to life.

“There are some teams that are ambitious and want to make a change. But we need to make sure that that it’s the right thing to do, and that we’re looking at it as a strategic initiative . .  . we don’t want to break something that’s working.”

Angel-Lalanne agreed, adding that when it’s done right, this kind of Voice of the Customer insight can positively transform organizational cultures.

“We now have partners coming to me asking, ‘Oh, can we listen to some customers? Can we ask customers this?’ And that’s been a goal of mine since I’ve been in my role for six and a half years, since I first joined this this organization,” he said. “I’m happy to see that momentum.”

View Comments (0)

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Scroll To Top