Banks use AI, taxonomy and machine learning to identify the causes of call center complaints and create a more seamless client journey.

“If you’re not identifying all of your complaints up front, then how can you run the analytics, draw themes, do root-cause analysis and stop the complaints from happening, or at least reduce the complaints from happening,” Nicole Wadlinger, head of business management and governance of the consumer bank at TD Bank, said earlier this week in Las Vegas at CBA Live.

That’s where machine learning (ML) comes in, Scott Hamilton, banking strategy executive at Prodigal Technologies, said. In the past, the bank could assume why a client might have called the call center, but now through data analytics and ML “you can verify that’s why they called.”

Implementing technology that can understand the human element behind a complaint is key, Hamilton said. “The consensus in the industry is the only way to unlock human levels of automated accuracy … is using machine learning.”

For example, machine learning can understand human tonality, Wadlinger said. If a client called and sounded upset, that tone would be recognized by the technology and that call would be categorized as a dissatisfied customer. That would narrow down the reason behind the call so the root analysis can begin.

In verifying the reason for a client call, whether it’s a fraud dispute or declined debit transaction, banks can then fix the end-to-end client journey from the root of the problem, she said.

“Root cause is so critical because that’s what you need to get to at the end of the day,” she said. “You understand the root cause analysis; you understand what you need to do to fix it.”