Customer experience improvement remains a top initiative for business and information technology leaders.
The stakes are high in CX. My research has found that 95% of companies now compete on CX and, last year, two-thirds of millennials admitted to switching brands because of a single bad experience. Contact centers and the customer experience professionals who staff them are the unsung heroes of many businesses. However, most organizations find it challenging to have well-trained professionals available to handle all customer inquiries.
The amount of information CX teams must master (or be able to pull up instantly) to respond effectively to customer inquiries and complaints is staggering. That’s why it’s no surprise that one of the biggest words in CX today has just two letters: AI. More companies across all industries are turning to artificial intelligence solutions to beef up their CX capabilities, reduce response times and alleviate agent burnout.
AI for customer contact centers was front and center at NICE Interactions 2024, the company’s user group meeting in Las Vegas last week. The maker of the Enlighten Copilot AI companion for customer contact employees pulled back the curtain and had two top customers explain how AI technology has transformed their contact centers, benefitting customers and staff members.
AI for self-service automation at Sony
Roger Brewer (pictured) cut his teeth fixing Walkman stereos, camcorders and other Sony products in Arizona. He rose to run the company’s customer excellence center for a decade. Now, he’s director of service tools and technology for Sony Electronics. Brewer walked through the successful implementation of Enlighten Copilot to empower Sony’s CX team.
He described how the Sony CX team evolved from relying on binders of information about the company’s products to an AI-enabled solution that gives the staff fast access to the accurate information they need to respond quickly and effectively.
“We wanted a balance of voice — people our customers could talk to for answers to their questions — and self-service opportunities that we could provide with increased operational efficiency,” said Brewer. Another goal was to measure the effectiveness and ROI of self-service opportunities. “We wanted to go where no one was going before,” he said. “We wanted to provide an omnichannel solution in a seamless operation.”
After talking with several vendors, Brewer and his team put the top contenders, including NICE, to the test. “We gave NICE about 60,000 CX chat records to see what they could do with them,” he recalled. Then Enlighten Copilot got to work. “NICE said up to 40% of those chats could have been automated. That was the green light we needed to move forward. Enlighten Copilot was the end-to-end solution that took the guesswork out of CX automation with its ability to analyze conversations, find the best ones and create a flow for automation so we didn’t need agent staff” for every call, said Brewer.
From a best-practices perspective, Brewer recommended that organizations considering AI for their CX operations start with analytics to see customers’ intents when they contact your company. “Look at the details. What are customers asking? What are they looking for when calling about their products? Trying to read through hundreds of thousands of interactions to find the ones that work and those that don’t and drill down to see how agents resolved them would take years. Enlighten Autopilot did it in minutes.”
Brewer said Sony took all those customer insights and “built and tested and retested and tweaked until we had confidence that we had a solution that would perform well.” Sony implemented Copilot in stages. “We launched in voice first, with a handful of intents last September,” he said, “In October, we added it to the chat solution. Then, in November and December, we added more intents. And in January, we added it to our chat channel. And since then, we’ve been fine-tuning and tweaking, making the product work even better.”
Lessons learned
“One of the things we learned is we should have involved our subject-matter experts earlier in the process,” Brewer said. “Get experts into the conversation. We looked at all the ways agents were handling customer questions. Initially, you won’t have an intent for every possible customer question.”
He noted that “there will be some frustration from customers who don’t want to talk to a bot. So, build a way they can get to an agent. And set expectations for every stakeholder, including the AI solution team. It will be a learning curve for them, too.”
Next steps for Sony
“We’re going to continue to fine-tune those automation and look at the output of Enlighten Copilot to find new opportunities for automation,” Brewer said. “We will look at customer feedback to see what customers say about these automations. And, finally, we’ll look at the knowledge solution expert because we believe that’s the next step for automation.”
Back to the future for a leading window treatment maker
The window shade and blinds business is different from that of consumer electronics, but Hunter Douglas’s implementation of AI for its customer experience teams followed a similar path to Sony’s.
“When I was a customer service rep for the company in the 1990s, we had much more dynamic conversations with customers, a better flow back and forth,” said Jeremy Markey, director of workforce experience at Hunter Douglas. “Since then, we’ve had a lot of innovations and solved a lot of problems. And we have a ton more data about our customers.
That, he said, enabled the company to provide better, more uniform experiences and find the interactions they can handle independently, leaving the most complex interactions to customer service representatives.
“We’ve added a lot of technical depth to those CSRs, and we found they have between 40% and 50% dead air when they talk with a customer,” Markey said. “They’re not twiddling their thumbs or chatting with friends. They’re clicking all over their computer, trying to find the knowledge they need to give the right answer.”
Because the products Hunter Douglas sells are customized, the company has built a formidable library of more than 3,500 knowledge articles over its long history. “Some are just a page or two, but some are over 60 pages,” said Markey. “Some directions will take the CSR to the right knowledge article, but that results in dead air. The customers hear typing, waiting for the rep to say something. That’s a lot of pressure on the CSRs.”
Enter AI
When Hunter Douglas started searching for an AI-based CX solution, it wanted one that would enable a CSR to ask one question and get the correct answer — no more reading through an article to find what they needed.
“We put together vendor evaluation criteria,” Markey explained. “The first was ‘one question, one answer.’ I needed something ready for the off-the-shelf contact center on day one.”
Hunter Douglas started its AI search with 26 vendors, which became three finalists. The final test was what Markey called “a 246-document gauntlet.”
“We gave each vendor all the documents and one week to work on them” to apply their AI capabilities. Markey said the company held in-person meetings with each finalist and asked each 20 questions to see how well the AI solution could respond. “The first vendor got one out of 20 right. The second vendor provided correct responses on six out of 20 questions. But the third vendor got 19 out of 20 questions correct.” That was NICE.
“Even the one question that NICE didn’t answer perfectly, they could show us the document with the answer,” he added. “We’re working to train Copilot to read the answer without the need to restructure our data.”
Lessons learned
Finding the right AI solution for a customer contact center is time-consuming. But achieving the goal of fast, accurate responses to customer questions and problems is worth it. For customers who prefer that method, using AI to augment conversations with CSRs is an approach many organizations should consider as they look to add AI to their customer contact centers.
