Whither the Agent Amid AI's Rising Popularity?

Customer service has a big problem, Well actually, it's about 15 million problems. They are called agents, and they represent more than three quarters of the cost of customer service at most companies worldwide.

That is a lot of motivation to try technology: interactive voice response systems, chatbots, call routing, workforce management, quality management, digital customer engagement. The list of technology deployed in the contact center goes on. Most of this technology is designed to make agents more efficient and productive or to allow customers to self-serve, avoiding the need for agent interactions altogether. Ironically, all this technology has not reduced the number of agents in the world. The number has held steady at somewhere between 15 million and 17 million for a decade or more.

How can this be if companies keep adding new technology to drive efficiency? Well, there are a few answers. Digital channels opened service to users who previously just got lost on the website and would not have bothered to call. This is good for the company, but it drives a need for more agents, not fewer. Self-service applications are often poorly designed, frustrating customers and disappointing companies with low usage. The advent of cellphones led to many more consumers who need support and have the means to ask for it.

Artificial intelligence, and now generative AI, is the next hot technology that companies are using to increase automation and agent efficiency. According to a recent study by McKinsey, the top three business uses for AI are customer service (26 percent), personalization (23 percent), and customer acquisition (22 percent). All three of these have value in the contact center. The next wave of solving our people problem with technology is here

It's hard to look at generative AI and and not think that the number of agents needed isn't going to drop significantly. Consider the following:

  • Generative AI-driven chatbot and voicebot applications will take about a tenth of the programming that is required for chatbots driven by natural language understanding and machine learning capabilities.
  • The conversational ability of generative AI will make it natural, easy, and efficient for customers to use for self-service.
  • Generative AI is already increasing agent efficiency by taking over post-call summarization work and will automate much more of the drudge background work associated with customer service.
  • Next-best actions and suggestions for agents will drive a new level of agent efficiency and success.

Given the above, I would expect the number of agents worldwide to start to drop significantly and quickly. I believe we will see a reduction from 15 million agents to around 10 million in the next two to five years, and for that number to keep dropping after that. Many more simple interactions will be handled by self-service, agents will be able to solve more problems faster. The job is not going away. Agents will be needed for at least the next several decades, but the job will get more complex. Agents will be dealing with questions requiring delicacy, complex situations that require deep insights, or interactions where things are difficult.

An example of people's preference for working with humans for complex interactions can be found in a recent Forrester survey. Of the 508 consumers surveyed, 71 percent preferred self-service when making a payment, compared to 29 percent who preferred a human for the same task. In the same survey, 91 percent of respondents preferred to interact with a human to discuss complex financial information. The more complex the issue, the more people want to speak with someone.

I don't know if my expectation will be correct. This is a much larger technology shift than we have ever seen before. But the agent's job will change for sure. I'm less clear about how many agents we'll continue to need. We have 15 million agents today because people provide value in customer service, and that is not going to change. >

Will companies ride this new level of technology to provide a new level of service? Imagine a world without hold times, imagine agents who are dedicated to really providing value. Maybe this new technology provides enough cross-sell and upsell potential that contact centers become profit centers and agents who generate revenue will grow as they provide more value than ever before. It's a stretch, but we can dream, can't we?

Whither the agent? I dunno. I can tell you the job will be harder and more professional. My gut says the number of agents will be reduced, but I would not be surprised to see far more agents than we expect. We've been holding at 15 million agents for quite some time now, and I'm suspicious that generative AI will be limited in its ability to change the direction of that trend.

Max Ball is a principal analyst at Forrester Research.