The drive to contain agent costs is nothing new in the contact center. We’ve seen some fairly extreme examples lately, such as Frontier Airlines' late 2022 announcement that it was discontinuing phone-based agent assistance altogether. And now, the rise of generative AI systems such as ChatGPT has renewed the temptation by some to envision a more customer-facing role for AI in handling customer service.
To be clear, I’m not hearing any voices calling for generative AI chatbots to supplant agents in any kind of flash-cut. But there will be a conversation about how far to push AI in replacing agents’ functions. That’s one of the reasons I’m excited about the “Disruptive Dialogue” I’ll conduct with Dan Miller of Opus Research, who will join me on the main stage at Enterprise Connect 2023 on Wednesday, March 29 for a conversation on generative AI.
Miller is one of the leading authorities on conversational AI, and he’ll have some great insights to share about the role that generative AI tools like OpenAI’s ChatGPT and Google’s Bard will have on contact centers, as well as on enterprise communications overall. I’m not going to steal any thunder here, but when I chatted with Miller this week to prepare for the session, I came away even more convinced of the urgency behind enterprises understanding and developing a strategy for leveraging generative AI.
The usual caveats apply in any conversation about generative AI: Concerns about accuracy, compliance, and security have to remain top-of-mind. But as the AI gets better, I wonder: Will the discussions spend more time on whether or not people fundamentally want to interact with AI?
In a session she’ll be doing at Enterprise Connect 2023, Robin Gareiss of Metrigy may dive into some data about what types of interaction produce the highest customer satisfaction scores. Almost 39% of those Metrigy surveyed say voice phone calls are still the top channel for customer satisfaction. Another 17% say textual interaction with a live agent produces the highest score on this metric, while 11% cited video calls. On the other hand, 21% said self-service provides the highest satisfaction, and another 11% cited textual interaction with a virtual assistant.
So at this point, you’ve got about two-thirds of a customer base preferring interaction with a live agent, and one-third aligned with not interacting with an agent (via either self-guided or virtual assistant). The nearest opportunity for gain would be in that 11% who see virtual assistants as providing the highest customer satisfaction. Presumably, generative AI can keep those virtual assistants continually improving and taking on more service opportunities, while allowing for a measured, conservative approach that doesn’t expose the enterprise to the potentially serious effects of “bad” or flawed AI directly delivering customer service.
Clearly, generative AI can do more than make chatbots better. It can help agents deliver better customer satisfaction as well. That’s important as agents are asked to take on more complex and critical tasks—they need the best tools the contact center can give them.
In addition to our Disruptive Dialogue, Miller will also lead a session on Conversational AI in which two large-enterprise users, from Lowe’s and Johnson & Johnson, will share their experiences and advice. I’m excited for this and for all of our AI content, as well as our sessions on hybrid work issues, E911, and everything else your enterprise IT/communications organization is dealing with. I hope you can join us in Orlando!
