Customer service and the generative AI advantage

Customer service is the proving ground for generative AI

Customer service is the tip of the spear for generative AI—the function that pierces through the unknown to deliver unprecedented business value. In fact, as this new technology disrupts how work is done across the enterprise, customer service has become the C-suite’s top priority for adoption. That’s no surprise, as it’s the next logical step for companies that have already been using traditional artificial intelligence in customer service for years.

From chatting with customers to creating targeted content to optimizing call center and contact center performance, generative AI is taking the transformation of customer service to the next level. By enabling dynamic and personalized experiences for both customers and human agents, it has the potential to supercharge traditional AI, spurring a seismic shift in productivity and effectiveness. Making the right bets can pay off exponentially—but where companies should invest depends on where they’re starting from.

Every leader we surveyed says their organization plans to use generative AI in customer service—and 67% say they’ve already begun.

So, where do business leaders at different stages in the AI journey see the most promise? To answer this question, the IBM Institute for Business Value (IBM IBV) surveyed nearly 1,500 customer service managers, directors, and executives from organizations that have used conversational AI for at least 12 months across 34 countries and all major industries. We asked how their organizations are using generative AI in customer service today, which use cases show the greatest potential, and where this technology is already delivering the most business value through automation and augmentation.

Overall, customer service leaders agree that adopting generative AI is essential for their business. In fact, every single respondent says their organization plans to use generative AI in customer service—and 67% say they’ve already begun. More than half (54%) of these organizations have deployed generative AI in one to four customer service use cases.

But not every organization plans to use generative AI the same way. Our data suggests the number of years an organization has used conversational AI, which is designed to understand and respond to customer queries in natural language, is a telling predictor of whether it will be an aggressive early adopter of generative AI. We see that organizations with the most experience using conversational AI in customer interactions have the confidence to be bold, implementing more sophisticated use cases. For instance, 89% of organizations that have used conversational AI in customer service for at least three years are already using generative AI to answer customer queries directly.

65% of customer service leaders expect using generative AI with conversational AI to increase customer satisfaction.

However, our research also reveals that using generative AI in conjunction with conversational AI can improve customer experience, support agents, and deliver significant business benefits, regardless of how much AI experience an organization has. Veterans do see best-in-class performance—but novices could gain the biggest edge over their peers. This means organizations at every maturity level have opportunities to provide better customer support, outpace their competitors, increase customer engagement—and deliver game-changing performance improvements. They just need to know the best next step to take. 

AI gets a nitro boost

When conversational AI came on the scene, it helped companies improve on early chatbot experiences, which were driven by rule-based systems that delivered pre-defined responses. These early AI systems were used primarily to address common, easy-to-answer customer questions and had limited capabilities.

Conversational AI improved on the chatbot experience by leveraging natural language processing (NLP) and machine learning algorithms to understand and respond to customer questions. When trained well, these AI models sound more like humans and less like machines. However, while these enhanced AI assistants can successfully execute much more complex interactions, their capabilities eventually hit a wall.

Generative AI offers the next evolution of AI technology. Using natural language generation and customer data, it answers customer questions with more fluent, contextually relevant responses. It can also tap into a customer’s interaction history to tailor responses and deliver more personalized responses. These capabilities let customers chat with generative AI assistants in the same way they would engage a human agent.

More than 40% of organizations are using generative AI to create test cases for training conversational AI.

What’s more, the applications of generative AI go far beyond direct interactions with customers. This technology can enhance the customer service function more generally by supporting customer service agent training, increasing personalization, translating content, and predicting future customer behavior to increase agent productivity. It can also support customer-facing conversational AI by generating test cases and dialogue, as well as reviewing interactions to identify opportunities for improvement.

Experience matters—but generative AI is a tide that raises all boats

Our research shows that, on average, all organizations using generative AI in customer service report higher levels of customer satisfaction than those that don’t. However, organizations that have been using conversational AI longer report the best business outcomes overall.

A bold step forward: Veterans are experimenting with more customer-facing use cases than novices.

Download the report to learn how organizations are leveraging generative AI in customer service today, how it can impact key performance metrics—including cost per contact and ROI—and which approaches work best for novices and veterans. Then explore an action guide that outlines how organizations in each group can optimize gen AI to get the most value.