What Callers Often Want
Will automation kill human phone support? The short answer is not for a while.
In a previous role our executives worked on developing voice virtual agents. They built solutions used today by some of the largest call centers, including ones handling over a hundred million customer calls a year.
From their experience often about 50% of callers in the US wanted to speak to a human agent. This rate varied per industry, for example at telecom companies often 70% of callers end up speaking to human agents, while at banks only about 30% end up speaking to human agents.
When callers are experiencing an issue and take the time to call customer support they often want a person to hear them out, offer them sympathy, along with a solution. Sometimes they also want to haggle with a person and get a special deal.
What Callers Often Say
Once customers connected to phone agents they often didn’t not know how to describe their problem. They would hint at symptoms they were experiencing, and would struggle to find the right words to describe accurately what issue they had.
Those same customers sometimes ended up telling stories as part of describing their issue. For example they might have told a story of how a relative wronged them and added themselves to an account without asking them. Often seeking sympathy and corrective action.
Other times customers engaged in small talk. They would want to connect with the agent, and would ask questions about where they live, what sports team they like, or questions about the weather.
Comprehensive Call Automation Difficult For Now
Given these myriad of needs and expectations callers have when contacting customer support, it is currently challenging, even using the latest large language models (LLMs), to reliably automate conversations in a way enterprises can trust. Enterprises after all need deterministic outcomes that follow their guidelines.
At the same time, callers need the ability to talk around a problem, to tell stories, and to engage in small talk. They sometimes want empathy, and other times they might want to get “good deal”. So while voice automation can work for simpler interactions, it can be challenging to implement otherwise given the gap between the needs of callers & enterprises and the current state of the art voice automation solutions.
