CX Leaders Share Inside Look at Their AI Journeys

The Gist

• Cautious integration. Many leaders are taking a measured approach to AI integration to ensure readiness. 
• Agent efficiency. AI is improving agent-customer interactions by handling mundane tasks, freeing agents to focus on more valuable activities.
• Experimentation and impact. CX leaders are experimenting with AI while being mindful of its impacts on both customers and employees.

LAS VEGAS — The biggest theme at Customer Contact Week here at Caesar’s Forum is artificial intelligence. It’s inescapable. Uses for agents, for customers, for managing and analyzing vast amounts of data. It’s here to stay. But not everyone is at the same stage of integration. 

While vendors tout fancy solutions and promise everything under the sun, many practitioners are taking a measured approach, considering exactly how and where AI makes the most sense for them.

We caught up with a few customer experience and contact center leaders here this week to find out exactly how they’re approaching AI. (Editor’s note: check out our other coverage from Customer Contact Week.)

Taking a Cautious Approach to AI 

Nadine Cotter, AVP, client services at Canadian Western Bank, said her company is at a younger stage, with a smaller contact center and a focus on expanding its channels. With AI, the plan is to time its entry so that it doesn’t feel behind. 

The main difficulty lies in knowing where to begin, particularly if you lack a person in your routine who is somewhat knowledgeable about the subject, she explained. “You’re kind of leaning into your gut and whatever research and data that you can get your own hands on in terms of how to go forward into that space cautiously,” she added.

Cotter continued that she’d like to take a cautious approach and not swim upstream too fast — launching too early just for the sheer excitement or testing it and feeling like you’re behind already.

“I just want to temper that with just being sure that it’s not only at the right time, but we’re fully prepared for it, we’ve tested it, we feel really confident in it,” she said. “But also embracing perhaps some mistakes and learning along the way and iterating and evolving.”

We come with no experience, said Cotter — which is part of why she’s at Customer Contact Week. But, she said, she recognizes the need to start dipping their toes into it so that it’s ready at the right time.

Related Article: Effective AI Implementation Starts Here

Using AI to Improve the Agent-Customer Experience

Pasquale DeMaio, VP of customer experience services at Amazon Web Services (AWS), which offers contact center software, said that when it comes to integrating AI into the contact center, have AI do what AI is good at and let people do what people are good at.

For agents, one big area AI can help is in after-call work, something with which some agents struggle. However, DeMaio added, they weren’t hired for their writing abilities — they were hired because they’re amazing at talking to people and empathizing and understanding. “And so this takes that off their plate and lets them spend that time doing something else. It’s much more valuable,” he said.

It’s a simple but powerful thing, he continued. “If you can save 30 seconds of summary writing at the end of every contact for many customers, that’s millions and millions of dollars, and the agents are not doing boring work that they hate. So big win-win.”

Another big area of focus that sees real results, he said, is agent assistance. Previously, AWS had shown agents advice and snippets from wikis to help them during calls. But there was a lot of cognitive load of getting those tips outside the context of the conversation.

Now, the technology can understand the conversations being had and make recommendations in ways that are useful for the agent. According to DeMaio, some AWS customers have been able to save 10% off their handle times with this capability.

Agents choose their jobs because they love helping people, said DeMaio. They love working with people, even though it can be an exhausting job.

“I go sit with them and listen in and sometimes I’ll even take the calls,” he said. “And I’m just blown away by the amount of passion they bring to that. And so I feel like if they’re going to put that level of passion into the work, we should make it a great experience.” And it’s not a coincidence, he added, that it ends up with better outcomes for customers, too.

View all

Experimenting With AI and Employee Changes

Heather Day, GM, customer experience strategy at Progressive Insurance, said her plans for the year for AI center around experimentation.

To advance predictive AI foundations, they should leverage emerging technologies while cautiously considering the downstream impacts on both customers and employees, she added. It’s about wanting to be very bold in terms of the opportunities that are out there, she explained, while also being measured in terms of how they go about it.

When it comes to talking with employees, Day said, they’re looking at the opportunities that AI will introduce.

“It may change the work that an employee is doing,” she said. “So it is more about flexibility in terms of what you want to do versus threat.” For the employees that aren’t willing to make that transition, she added, the challenges that technology introduces are always tougher.

Related Article: AI Drives Unified Customer and Employee Experiences

Enhancing the Human Experience With AI

Robin Jurkowski, a CX and contact center leader, said that she wants to find out how AI can help — not by replacing the human experience, but by enhancing it.

“I’ve used AI in the past for training, coaching, even motivating,” said Jurkowski, adding that she sees the technology as a way to create fun in the contact center. “So how can I use it internally and externally with my future customers to give them the best but yet the most human experience at the same time?”

Right now, Jurkowski explained, when you talk to AI, you usually know it. But she’s interested in the technology that performs at a level where, 99% of the time, the customer doesn’t know they’re talking to AI.

But if you can’t do it well, she added, it’s better to just acknowledge it. She pointed to the example of an automated system where the caller can hear faux “typing” in the background to simulate the bot looking up information. “It just seems very inauthentic, and like I’m trying to trick you and also think you’re so stupid that you don’t realize that the stuttering and the typing is completely false,” she said.

If you can do it well in a way where it sounds conversational, however? That’s something Jurkowski said she hopes is possible. But, she added, “I haven’t yet experienced it.” 

The Use Case-Based Approach to AI

Paul Herman, SVP of global digital transformation at Sprinklr, said that AI can mean very different things to different people. At one end of the spectrum, you have AI: absolutely incredible. At the other end, AI: awful idea. “The reality is somewhere in between.”

“Now,” Herman explained, “once people decide to do something, most of the companies doing a really good job are going at it very pragmatic use case-based.” And the people just turning it on and seeing what happens are usually the ones with the scary outcomes.

At his company, said Herman, it has proprietary AI models that have been around for a while — models that are focused on verticalized, customized models. And then there’s the power of generative AI. And the best path is to combine those models. “So the whole idea is, instead of having an unbounded dataset, producing an unpredictable outcome, how can we bound the dataset to produce a predictable outcome?” he asked.

Another big discussion, he said, is build versus buy. “My thing, honestly, that gets me concerned, is when I’m here at CCW, everybody is touting their own AI model.” It’s AI on top of AI on top of an AI, he said. And they’re going to compete with each other. “And who’s going to lose? You’re going to lose as a consumer,” he contends.

It will be a topic we’re all talking about in two years, added Herman.

“How do I get a handle on all my AI models? It’s gone nuts,” he said. “I think that’s going to be a big thing that people are going to start to figure out — how do we broker the AI model? And I think the only way to do it is to take the use case approach.”