Harnessing AI to redesign contact centre operations

By ,

Part one: Improving clock time and spotting emerging trends

Technological innovations have created the ‘customer monster’. Through no fault of their own customers have been given a wealth of communication tools to choose from and with so many organisations vying for a competitive edge, customers are been treated like royalty in attempts to win and keep their business. As a result, expectations have soared and they’ve become more demanding than ever.

Customers expect great service 24/7

Seventy four per cent of customers think that customer service departments should be open 24 hours a day, 7 days a week. That’s right, if they’ve got a question they expect an answer - and they expect it now. It doesn’t matter if it’s mid-morning, mid-day or the middle of the night, customers expect someone to be on hand to help, immediately.

Organisations must provide an omni-channel experience

This isn’t just on the phone either. Eighty three per cent believe companies should make it easier to be contacted by phone as well as webchat, email and social media. And if it’s a crisis they find themselves in, 52 per cent of customers want a well-trained employee right away. No time for triaging here.
Voice biometrics can drive down customer handling times

When a speedy response is critical, customers don’t have time for a plethora of security questions. It’s frustrating to be asked the name of you first pet when you think you’ve just become the victim of credit card fraud (and it’s even more frustrating when you can’t remember if it was Fluffy with a capital F or not).

Some banks have been ridding their customers of these headaches by using voice biometrics. Leading banks have seen the identification process reduce from 45 seconds to just 15 seconds when a caller is being authenticated through recorded voice prints.

Voice biometrics allows calls to be screened against a dynamically updated database of customer and fraudster voice prints in real-time, improving genuine customer experiences. At the same time, it frustrates fraudsters and deters them from calling the contact centre.

Deep learning software improves fraud detection

Keeping the criminals at bay isn’t an easy task and it’s not made any better by the never ending challenge of improving customer experience. It shouldn’t be a surprise to anyone to hear that customers don’t like being treated as criminals. But this is exactly what happens to some.

Banks have deployed a variety of different fraud detection schemes. But no matter how good the detection scheme is, there are always genuine customers who get pulled up on ‘suspicious behaviour’. Some banks have turned to AI to identify fraud and reduce the number of genuine customers being caught in the net.

Through the implementation of ‘deep software’ banks have moved the process of operational decisions from humans to AI. Using contrasting models, the deep learning systems compare two or more strategies and promote the one that performs better, to build a successful fraud detection roadmap. Champion-challenger methods apply similar logic to AB testing used by marketers to measure effectiveness of advertising. Banks following this approach have improved fraud detection by 50 per cent and cut the number of genuine customers getting caught up in the process by 60 per cent.

To keep reading about how artificial intelligence can augment humans in the workplace and bring more meaning to our professional lives, download our latest whitepaper here.  


Joseph Walsh


Let's stay in touch