The Influence of Escalation

As someone who both studied statistics and worked as an AT&T market researcher for several years before becoming an industry analyst, I know that the responses to survey questions are dependent on the wording of the questions. It is easy for me to not just imagine but to calculate how different the data would be if the choices had been limited to the other four: open to it, maybe, not sure, or not at all.

Removing the “escalate” option and recalculating the responses — making the big assumption that everyone who chose it would distribute their responses similar to those who did make other choices — presents a very different picture. Instead of a modern view of today’s consumers — willing to try new technology as long as they have backup to talk to a human if they get stuck — only 13% of respondents would have reported that consumers are open to chatbots. Over 50% would have answered maybe, and a full 27% would have said consumers would not at all be willing to use chatbots.

Page from the History Books

The lesson of our story: You can draw a correlation between the deployment of chatbots today and the IVR in the 1990s. There are ways to implement the technology that serve consumers well, that give them faster answers to easy questions. It is also easy to fall into a trap of creating a new kind of IVR jail, where consumers are unable to get the human help they need when they need it.

What won’t happen in 2020? Companies are not going to abandon human agents in favor of chatbots. But they are going to continue to deploy them for simple interactions and work to make it easy for consumers to reach out to a human agent when required.