Generative AI's Contributions to Contact Centers and Customer Service

Generative artificial intelligence has embarked on an amazing journey, yet no one knows with any certainty where it is going, what will come next, or how it will end.

What is known is that most contact center and customer service technology vendors want to be part of it and are augmenting existing applications with these technologies, developing new generative AI-based solutions, and embedding them into their platforms' foundations.

The contact center technology market is experiencing massive disruption and innovation as generative AI alters short- and long-term product roadmaps to enable vendors to harness its power and catapult their offerings into the future.

Companies are also excited about generative AI's potential, when properly leveraged, to fuel the next wave of automation and productivity improvements. As generative AI picks up tremendous momentum throughout the technology community, there are growing concerns about its impact on security and privacy. While guardrails and regulation are needed for this new technology because of its power and intelligence to generate human-like content, it is here to stay.

Still, contact centers and customer service departments remain people-intensive, despite the digital transformation. While digital interactions and inquiries are growing, phone calls are not going away. This increases the need for live sales, collections, and customer service employees, especially as companies scale and expand. Self-service has become consumers' preferred service channel, and technology vendors have introduced applications to help managers improve productivity, yet many companies still cannot hire and retain the necessary caliber of employees to properly staff these mission-critical customer-facing functions. Just as the service world reached this critical juncture when many companies felt compelled to compromise service quality, generative AI entered the market and set the stage for a new service era.

The most common (and well-known) use of generative AI in service organizations is to create text-based outputs, such as emails, real-time agent guidance, post-interaction summarizations, and conversational responses to self-service inquiries. Generative AI is also being used to produce more accurate call transcripts, pinpoint internal company issues, identify customer needs and wants, find top trends and challenges in customer journeys, etc.

Due to its ability to create content based on patterns found in large language models (LLMs), applications for generative AI in contact centers and customer service organizations are widespread and growing. Consider the case where generative AI is leveraged to present an agent with the steps to properly resolve an esoteric procedure. If the employee had to find and read a lengthy knowledge article, it could take an excessive amount of time to address the inquiry. With input and direction from generative AI, the agent is presented with just the right amount of information to get the job done accurately and on a timely basis. This highly practical and real use of generative AI demonstrates why it is ideal for service organizations; however, these solutions are reliant on having access to LLMs that contain the appropriate, targeted, and tagged data needed for each department.

The Risks of Generative AI

Generative AI has captured the imagination of executives and service leaders due to its ability to create and render content that is generally as accurate and appropriate as would be provided by a live agent. To accomplish this and perform similar tasks, generative AI applications need access to LLMs that can consist of millions and possibly billions of text data points or other relevant information, to find patterns on which to base its generated content. This is both the strength and weakness of these solutions, as they are only as accurate as their training data and therefore will carry forward inaccuracies and biases found in the LLMs.

These solutions are also known for generating hallucinations when they don't have enough information or the right data on which to base an answer. But while this is a risk that companies want to avoid, they face the same concerns from live agents who could give an incorrect answer when they don't have access to accurate responses.

Security, privacy, and anonymization are other major areas of concern when it comes to LLMs. Open-source generative AI providers generally deliver access to an extremely broad data set. While these vendors invest substantial effort to ensure their LLMs do not contain inappropriate or damaging data, the broader the data set, the more likely it is to include information that could not be appropriate for some audiences. There is also the issue of these LLMs picking up patent-protected content that intellectual property (IP) owners would not like to see replicated.

For these reasons and more, many contact center and customer service technology vendors and companies are creating their own LLMs that they can target, tag, and curate. Entities like commercial banks, insurance companies, and healthcare organizations are increasingly looking for verticalized training data that is relevant for their specific uses.

When it comes to generative AI, the market is limited only by our imaginations and programmers' ability to code these solutions. While LLMs are a gating factor—and appropriately so—for most organizations, there are already ways to address this issue, one of which is to access the company's massive contact center data repositories as a primary source for training data.

Many contact center and customer service technology vendors have already figured out how to augment their solutions with generative AI, but that is just a first step to get them into the market. The next phase is to incorporate generative AI technology more fully into the foundation of their platforms and solutions, making it an essential workflow element to further automate tasks that previously required manual intervention. While this can benefit service organizations and dramatically improve the CX, EX, and productivity, it is just the beginning of what generative AI and other AI technologies are expected to contribute to contact centers and customer service organizations during the next few years.

Donna Fluss, president of DMG Consulting, is an expert on contact centers, analytics, and back-office technology. She has 30 years of experience helping organizations build contact centers and back-office operating environments and assisting vendors to deliver competitive solutions. She can be reached at