NBN’s AI strategy is moving to an agent stage where network wholesalers can expose their data to retailers and increase retailers’ visibility into their networks.
Richard Halliday, the national network construction company’s general manager of operations, told attendees at the AWS Summit in Sydney that NBN had previously taken a slow approach to AI technology, cautious about releasing large amounts of data.
However, he said NBN Co was now ready to start looking at agent applications and the ability to give retailers greater access to their network data was emerging as a promising business case for the technology.
“We’re at the stage now where we’re trying to get it moving.” [AI strategy] We’re looking at how we can move to agents and be predictive, how we can look at the network and find faults and do that automatically and effectively,” Halliday said.
As a result, NBN may soon share more information about its network with retailers, he said.
“We don’t do things like that unless we’re careful about how we approach them…But what this shows is that we… [network] “So how can we leverage the ability to expose more network data to retailers and the broader industry to bring the entire industry along on that journey?”
Mr Halliday also revealed that NBN, as one of the country’s largest providers of critical infrastructure, is exploring ways to use the technology to improve network resilience.
“Our goal is to make networks more predictive, self-healing, and adaptive, and we believe that using AI in these environments will enable us to do that, both to serve our customers and through our retail providers to end customers across the country,” Halliday said.
NBN Co remains cautious in its approach to AI given the regulatory environment in which network builders operate.
Halliday said this means the company is taking a “dual-speed” approach to AI, seeking to balance innovating quickly while minimizing risk to the nation’s critical infrastructure.
“We have to be very conscious about how we expose what we do to the industry at large, and we also have to be very conscious about: [taking] “We need a dual-speed type of approach to how we move forward with AI development, where the core capabilities need to remain highly resilient and capable, while much more innovation can be done at the edge,” he said.
“We’re looking at how to actually manage and balance these two functions.”
