Few things will cause a customer to immediately request — or demand — a live agent like a bot that misunderstands them.

Few things will cause a customer to immediately request — demand, even — a live agent like a bot that misunderstands them. A CX leader implementing AI for self-service should first understand why that can happen and then take several steps to eliminate that friction.

“A bot can understand every word a customer says and still completely miss the point,” Brad Cleveland, CX expert and author, told sister publication CX Dive.

This is because customers don’t speak in neat scripts, may conflate multiple separate issues into a single statement or may use sarcasm or irony to convey the opposite intent. Additionally, context can be missing, and the knowledge on which the bot draws its information can be outdated.

“This is the Achilles heel of AI and customer service across the board. It's a huge need to make sure the underlying knowledge is sound,” Cleveland said.

Organizations should also make sure the knowledge bases from which the bot will draw its responses are current and accurate. And, of course, testing is a critical step.

“Test with real customers and real conversations. And not just with a perfect, polite customer,” Cleveland said. “Test with frustrated customers, fast talkers, a person with dogs barking in the background or at a busy airport where they’re trying to get something done while waiting for a flight.”

Lastly, CX leaders should measure failures and make escalation to a human agent easy — which may even encourage AI adoption .

Source link