Dive Brief:
- Customers don't like to repeat themselves. A majority, 60%, say they will repeat themselves only once before abandoning an automated customer service experience, according to a Parloa study released Monday. Another 11% will leave the first time they are asked to repeat anything.
- Time is of the essence as well. More than half of respondents — 56% — say they will work within an automated system for less than three minutes before asking to speak with a person, according to the survey of 1,001 U.S. consumers.
- The top customer service pain point was “talking to a bot that doesn’t understand me,” followed by long hold times, multiple transfer and having to repeat information.
Dive Insight:
Customers don't want to be delayed by bots that don't understand them or force them to repeat details down the line if a call escalates. Their frustration is reflected in their lack of trust.
Only 14% of consumers say they completely trust future AI systems to handle complex customer service requests better than humans according to Parloa. Another 37% of respondents say they expect adding more automation to produce worse outcomes before better ones.
The distaste for automation carries over to IVR systems as well. Only 7% of respondents say IVR technology consistently resolves their issues, while 45% say IVR sometimes helps but rarely manages to achieve full resolution.
Creating automation that customers like is a matter of building technology around lessons from actual customer interactions rather than making assumptions about what consumers need, according to Julie Geller, principal research director at Info-Tech Research Group.
Companies can reduce the need for repeats and improve speed of resolution by training AI on the most common requests, according to Geller. Teams can dig into internal data and use that as a jumping off point to improve pain points.
“Organizations should analyze contact center transcripts and chat logs to identify the highest-friction issues customers encounter,” Geller said in an email. “We need to start treating friction as a system signal rather than just a metric.”
Customer service teams need to measure automated resolution quality as well as resolution rates, according to Geller. An AI assistant that can successfully handle very straightforward questions but struggles with even common exceptions can ultimately erode trust and confidence.
Many consumers are still open to automation when it works. Three-quarters of respondents say they would even prefer automation if it could anticipate their needs and act proactively, according to Parloa.
