AI Enhances Customer Service in Several Ways

Although the initial seeds of generative AI (GenAI) were planted in the 1960s, for the most part it didn’t come into full bloom until the introduction of ChatGPT in November 2022. Before that, GenAI had all but the most nascent use in business. As the technology entered the public consciousness, it spread across industries from entertainment to manufacturing to healthcare for a wide variety of uses, including customer service.

"At NICE, we see generative AI as a key component to enabling businesses to automate their customer service at scale,” said Carmit DiAndrea, director, product strategy and innovation: AI, analytics, and digital experience.

There are several different use cases for the technology, according to Heather Richards, Verint’s VP of GTM strategy. “When organizations first started to use it, it was being used almost as a catchall for all of the ways that you could automate anything. I think the approach that Verint has taken is to break down the customer experience into different workflows and tackling each one of those individually.”

Those who use AI in their customer experience solutions see the following as the primary use cases:

AUTO-SUMMARIZATION

Rather than an agent spending several minutes writing after-call summarizations, GenAI provides the summarization almost immediately after the call. Such summarizations provide important context for the next agent who will be interacting with the customer DiAndrea said. “They [also] provide supporting documentation for escalations and fulfillment and are an important datapoint for identification of trends and opportunities.”
NICE’s auto-summarization solution captures intents, actions, entities, outcomes, and sentiments as well, DiAndrea said. “We provide our customers with the ability to customize these summaries to capture things like part numbers or model numbers for specific products.”

“The customer experience spans everything from self-service through to assisted service,” Richards added. “It enables the agent to hit the ground running. If the agent can pick up the phone and immediately know what the customer has talked to the IVA [intelligent virtual assistant] about, what the customer has looked at on the website, what they’ve done previously… that’s automatically going to translate into a better customer experience.”

CONVERSATIONAL INTERACTIONS

“It used to be that you would have to programmatically tell the IVA word-for-word what it needed to say in response to a customer,” Richards explained. “You had to be very specific. GenAI enables that conversation to happen naturally. It knows what to say next. GenAI is generating the response to questions. That removes the requirement to actually have preprogrammed intents for every question or every interaction that your customer may have with your bot or intelligent virtual assistant.”

For example, before GenAI, if a customer asked a question about a product return, the IVA would have recognized the query was about a return. It would then need to ask what product was purchased and when it was purchased. Next was offering preprogrammed responses, Richards said. “Now you can use GenAI and a pre-curated list of FAQs. You instruct the bot to use this information to reply to the customer.”

Conversational AI provides these responses based on trusted knowledge sources such as website content and approved content that might reside in knowledgebases, documents, and other trusted sources, added Varsha Thalange, eGain VP of product management.

AGENT AUGMENTATION

Through a capability that many organizations call copilots, the technology offers immediate assistance to customer service agents. “It provides agents with accurate, relevant, personalized responses and knowledge guidance for customer inquiries,” DiAndrea said. “The solution reduces the cognitive load on agents while simultaneously improving customer experience.”

NICE customer Kaiser Permanente, with 94% of its agents using the augmentation solution, has experienced a 14% reduction in average call handling time within a year. An electronics manufacturer is successfully “containing” 30% of the transactions that would have previously been escalated to a supervisor with deeper knowledge about the customer interaction, according to DiAndrea.