Just before Enterprise Connect Orlando, I wrote a piece about the appeal of Artificial Intelligence (AI) in communications, especially for the contact center. What I saw and heard at the show, and since then, have only strengthened my opinion that AI is coming soon to a contact center near you. A recent Enterprise Connect webinar filled in some detail on these perspectives for me.
In the webinar, Zeus Kerravala of ZK Research presented his findings and projections about AI adoption in the contact center. Zeus found that AI was in involved in 35% of contact center customer interactions in 2018. He’s predicting steady growth for the next year and a half—42% by the end of this year, 51% in 2020—then a pretty quick ramp-up, to 67% in 2021 and 80% in 2022.
The farther out you go in making these kinds of projections, the less certain you can be, of course. Eighty percent of any market seems like a lot, but if AI really does find its best applications in automation of agent tasks—as opposed to eliminating agents—that seems like a foundation on which most contact center vendors would be able to build their systems.
Zeus presented data that, to me at least, reinforces the conclusion that AI won’t be, in and of itself, primarily a customer-facing feature, at least for voice interactions. The way he put it is that in the coming years, “Voice simultaneously becomes less important and more important” —less important because customers will continue to prefer self-service whenever possible (86% of customers use it today, according to Zeus), and more important because, when self-service fails to solve a problem or answer a question, live assistance becomes more of a make-or-break scenario. And live assistance will continue to include voice as the predominant channel; Zeus said 75% of customers use three or more channels (out of voice, social, chat, Web, and other), and voice is the only channel that appears in the top three across all age demographics.
It’s reasonable to assume that voice will continue to be the dominant channel for exception handling and problem resolution. Voice interactions will be the most complex and fraught encounters with your customers, so it’s critical that the agents who handle these calls are excellent representatives—and that they’re armed with the kind of information and insights that AI-driven back-end systems can provide. It’s also vital that they can use their time efficiently while avoiding burnout, a goal achieved in part by sparing them the kinds of mundane reporting tasks that AI can handle.
In the end, it matters because of one final statistic: According to Zeus, five years ago, 28% of companies said they competed on customer service; now that figure is 90%.
During our Q&A period in the webinar, which included Lori Stout, senior product marketing manager at Talkdesk, we discussed how to prepare for AI in the contact center. One key takeaway was that benchmarking and tracking metrics would become even more important. You obviously want to know whether scores for things like first call resolution improve as you implement AI-powered systems. But as Lori pointed out, the real yardstick for AI, unique among product features, is whether it continues to improve independently over time. Systems that give you a one-time boost are great, but AI is supposed to use the data generated by the system to create a positive feedback loop of continual improvement; otherwise the “I” portion of the AI is insufficient if not lacking.
So this is all a highly optimistic picture. It’s hard to make predictions, especially about the future, as they say. But AI is definitely coming to the contact center, and it’s not a development you can afford to ignore.