Artificial intelligence: Solving the little big problems

Published on the 20/08/2015 | Written by Donovan Jackson


artificial intelligence call centre interactive intelligence

Call centres are practically a defining feature of the modern world. They are also routinely a source of frustration, but AI could change that…

Notwithstanding the essential role call centres play in enabling cost-effective support and access, they don’t enjoy the most stellar of reputations. Oddly, perhaps, it tends to be the little things which cause the biggest annoyance and frustration.

Good news, then, that advances in applied artificial intelligence (AI) are increasingly making call centres more capable of accurately meeting requirements while delivering improved customer satisfaction.

That emerged in a discussion with Donald Brown, founder of Interactive Intelligence.

Starting out with a chuckle, Brown noted that AI “Has been on the treadmill of lofty premises and underwhelming delivery for a very long time. I specialised in it at grad school in the mid-1980s…but after that, for a long time, it became a dirty word.”

That’s because the premise of AI, while alluring, simply lacked the computing firepower and software maturity to actually do anything useful. “The predictions back then were for machine intelligence on a par with humans within a decade, but those hopes were dashed and AI faded into the woodwork.”

Like so many of the promises of the technology industry, Brown agreed, the initial overblown expectations just weren’t rooted in available capabilities. “Predicting the trends is easy; predicting the time frames is a bitch,” he laughed.

But, as routinely happens, time passes, Moore’s law took hold and an infinitely more capable technology ecosystem emerged.

“Though AI was pretty unfashionable for a time, we kept going with research and have seen considerable advantages which are now coming into the contact centre. Things like speech recognition, statistical analysis and helpdesk automation. In the last decade, and certainly in the past couple of years, AI is coming back with a vengeance, particularly around neural networks and machine learning. And it’s exciting to see that happen.”

Probably one of the biggest bugbears in the modern call centre is the process of authentication. It’s also pretty important, particularly when dealing with bank accounts or other sensitive information. “Speaker verification is one of the more interesting developments which is yet to come. Right now, it’s 20 questions. It’s a basic problem in which AI can help, identifying callers by analysing voice patterns and other contextual information,” said Brown.

The other area he highlights is related to speech recognition, already used to identify key phrases ‘in stream’, and which he says further drives the ability to address the common bugbears encounted with contact centres.

For example, how much easier or more pleasant would it your interaction be if you could consistently speak with the same person? Or, from a few phrases and other known information, such as previous calls and context from line-of-business applications, what if the contact centre anticipated your issue? And the agent answers the call proactively, ‘I believe your issue could be X, so I have prepared the following actions’?

What about technology that listens for certain words to trigger alerts? (For example, said Brown, “There are certain words which might indicate a heart attack – numbness in the left arm and jaw, shortness of breath – and those can kick start the necessary processes. That’s just one context, it could be used in any number of other ones.”)

Some of this sort of AI is already being put into place, said Brown. “Machine learning can really help reduce the frustration. In some cases, it can remove the need for person-to-person interaction altogether, if that’s what the customer wants. What inevitability happens is that simpler interactions become automated by smart software which can, in some cases, proactively deal with issues before a call to the contact centre is even necessary.”

And when a call is necessary, AI will increasingly be behind the ability for the agent to serve up a big helping of delight, rather than aggravation.

Post a comment or question...

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

MORE NEWS:

Processing...
Thank you! Your subscription has been confirmed. You'll hear from us soon.
Follow iStart to keep up to date with the latest news and views...
ErrorHere