Some types of artificial intelligence aren't ready for their close-up

Michael Ringman, CIO at TELUS International, has a unique vantage point on artificial intelligence and how the technologies under the AI umbrella are developing. He’s helping build an artificial intelligence competency for TELUS’s customer contact center and for its IT services company, which supports 30,000 team members globally. And, as part of the company’s digital transformation consulting group, he’s helping customers integrate various types of artificial intelligence into their companies.

Here, Ringman shares his thoughts on the types of artificial intelligence that are making an impact — and the types of artificial intelligence that still have a ways to go.

From your experience, what types of artificial intelligence are market-ready and what types of artificial intelligence have you found to be overhyped?

Michael Ringman: When you talk about types of artificial intelligence, there are a lot of different things that today get bundled into that term. I’ve seen everything from speech recognition and speech-to-text translation classified as artificial intelligence, all the way up to some types of data analytics and data mining in regards to trying to find things like voice of the customer.. It’s an exploding area across the board — whether you’re in customer service or you’re just providing IT services.

When you talk about speech-to-text, that’s a great example of where a lot of work has already been done. You see things like Google Translate out there, even Alexa and a number of the speech tools that are coming out with such as smart speakers that take natural language, convert that into a text, and then start to leverage back-end computing to do something based on the text. In the call center industry, for us, speech-to-text, voice analytics, voice of the customer — that has been hugely successful.

And overhyped AI?

One of the areas where AI is a bit overhyped, specifically in that customer service environment, is chat bots. Chatbots present a great opportunity — a huge way forward to simplify, to give more access to customer support. But I would equate chat bots to the voice activated [interactive voice response] cards of the early 2000s. IVR was going to change contact centers because it was all going to be handled by computers in the cloud. And you saw some really good implementations of speech IVR and you saw some really bad implementations of IVR. Fast-forward 18 years or so, we still have contact centers, and we still have a lot of different ways that people want to receive that service.

So looking at chat bots, a great place to get started is with simple, repetitive tasks for a customer. Things like changing a password. Even just acknowledging to the customer, “Hey, saw that you’re chatting with us. You’re in queue and we’re going to get back to you shortly. Is there anything I can help you with in the meantime?” So leveraging chatbots for simple, straightforward tasks up front can be helpful.

As they start to get more complex, providing support to a team member or an agent behind the scenes can add a lot of value to an organization — and won’t necessarily have potential brand damaging impact. Microsoft a few years ago released their chatbot out into the wild just to see what would happen, and while they said of those results, “Hey, we learned a lot out,” they were kind of disastrous.

So containing the chat bot so that your brand isn’t necessarily damaged but potentially improved by providing support to the contact center agent, you can start to train the those bots to understand more of what your business processes are without necessarily being directly in front of a customer where it could have potential adverse impacts.

Cloud Computing
Mobile Computing
Data Center
and ESG