Quick Overview

  • According to research released by Parloa on Monday, 60% of customers said that in automated customer service experiences, they would only repeat information once before giving up; another 11% would leave the first time they were asked to repeat information.
  • Time is also critical: a survey of 1,001 U.S. consumers showed that more than half (56%) of respondents said they would request a transfer to a human agent within less than 3 minutes of interacting with an automated system.
  • The most frustrating issue in customer service is "talking to a bot that doesn't understand me," followed by long wait times, multiple transfers, and having to repeat information.

Deep Insights

Customers do not want to be delayed by bots that do not understand them, nor do they want to be forced to repeat details downstream when a call is escalated. Their frustration is reflected in a lack of trust.

According to Parloa's data, only 14% of consumers said they fully trust that future AI systems can handle complex customer service requests better than humans. Another 37% of respondents expect that adding more automation will lead to worse outcomes before producing better ones.

This aversion to automation also extends to IVR systems. Only 7% of respondents said IVR technology can consistently resolve their issues, while 45% said IVR is sometimes helpful but rarely achieves full resolution.

Julie Geller, Chief Research Director at Info-Tech Research Group, said that the key to creating automation that customers like lies in building technology around lessons learned from real customer interactions, rather than making assumptions about consumer needs.

Geller noted that companies can reduce repetitive requests and improve resolution speed by training AI on the most common requests. Teams can dig into internal data and use it as a starting point to improve pain points.

"Organizations should analyze contact center recordings and chat logs to identify high-friction issues that customers encounter," Geller said in an email. "We need to start treating friction as a system signal, not just a metric."

Geller believes that customer service teams need to measure both the quality of automated resolution and the resolution rate. An AI assistant that successfully handles very simple issues but struggles with common exceptions will ultimately erode trust and confidence.

Many consumers remain open to automation when it works effectively. According to Parloa's data, three-quarters of respondents said they would even be more inclined to use automation if it could anticipate their needs and take proactive action.