Wireless network outages could be predicted and reduced, and customer service practices overhauled, if Canada’s telecom sector embraces artificial intelligence.
That was one of the messages from Telus Corp.’s T-T chief information officer Hesham Fahmy at the opening day of the Canadian Telecom Summit in Toronto on Monday, where discussions about AI applications and industry evolution are front and centre this week.
Fahmy said AI is useful for recognizing data patterns and detecting anomalies.
“If you think of productivity, every time there’s an issue, we were sending a truck out to a cell site or something to fix it. How much of that can you get ahead of before it becomes a crisis?” he said in an interview following a keynote address on using generative AI to increase productivity.
Mr. Fahmy’s speech acknowledged there are risks to added reliance on generative AI, including that the technology is geared to sometimes “hallucinate.”
“They are not designed to tell you, ‘I don’t know the answer.’ A researcher told me the best analogy to describe it is they are confident toddlers,” he told the hundreds of industry representatives in attendance.
“But that’s a big risk because then they can produce incorrect information.”
Jeffrey Maddox, president of Nokia Canada and chair of the Canadian Telecommunications Association board, said new technology often prompts the need for public reassurance surrounding privacy and trustworthiness.
He predicted AI trustworthiness and ethics would “earn special focus of regulation and policy” in the years ahead.
“Just look at the nervousness as we study the topic of AI,” Mr. Maddox said. “People naturally worry about the negative side of it and fail to see the immense upside potential that this technology could have.”
While caution is warranted, Mr. Fahmy said the industry should lean in and encourage “responsible AI” in its practices, as he believes any debate of whether there should be more integration of AI in the sector is over.
“What I can definitely say is this ship has sailed and this is happening at scale and the adoption will be very fast,” he said.
“How it’s going to materialize, what it’s going to be, I don’t know because we’re only seeing the tip of the iceberg but I can guarantee you there’s a world two, three years from now when almost every knowledge worker has got some element of AI in their workflow.”
When it comes to network outages and other behind-the-scenes technical malfunctions, Mr. Fahmy said Telus is testing whether generative AI trained on an “encyclopedia of huge post-incident reviews” can help solve problems more quickly when they pop up.
“It goes through all the stuff and says, ‘Well, it’s kind of similar to this incident that happened’ and gives you those suggested remediations you could do,” he said.
“I think it’ll be hugely powerful and productive for those engineering teams.”
Raising the level of transparency surrounding network outages is an ongoing focus of the federal telecommunications regulator.
The CRTC launched a consultation in March with the goal of developing a framework to improve the reliability and resiliency of telecommunications networks.
That study, which remains ongoing, proposes requiring Canadian mobile carriers to notify the CRTC, federal government and other relevant authorities when they experience outages. It would also force telecom companies to submit a post-outage report to the commission.
Mr. Fahmy said another potential application for AI relates to customer support, noting that companies such as Telus could lean more on technology to potentially reduce the amount of time customers are on hold waiting to speak with a telephone agent.
He said the idea could free up agents to spend more of their time and efforts ensuring customer satisfaction rather than trying to understand the problem a caller is experiencing.
Mr. Fahmy said Telus is “bullish” on the idea and expects increased use in the near term.
“No one loves to be on a hold for several minutes and then be passed from agent to agent. Well, what if you’re already kind of dealing with an AI agent right there?” he said.
“You’re actually helping customers get their outcome faster.”