TAL Australia is taking an “experimentation approach” to generative AI as it notes the emerging technology can have a “complementary and copilot role” to supporting staff.
General manager for technology delivery Hinesh Chauhan cited OpenAI research that 80 percent of jobs could incorporate generative AI, in no small part owing to “how easy this technology is to use”.
“This technology is quite powerful in terms of its adoption, and adaptation to different use cases,” Chauhan said at a Gartner Data and Analytics Summit in Sydney.
“Our focus here at the moment is experimenting. We’re looking at taking an experimentation approach.
“Anything else that’s much more bigger and large scale, guess what? By the time you’ve actually deployed the technology, it’s already changed.
“You have to be quite adaptive, both from experimentation in value pursued, but also from a governance perspective.”
Given the rapid uptake of the technology, “setting too rigid governance structures is not going to work either”, Chauhan added.
“We’re really curious about the technology and making sure our governance and policies are adaptive as well,” he said.
The nature of the insurance industry means there are people handling “large volumes of data”, particularly unstructured, which Chauhan said “is the sort of stuff generative AI can gobble up.”
“There are a fantastic number of use cases, but embedding it into the business process is also going to be important.”
Chauhan said the standard insurance journey “can be quite clunky”, with staff sometimes required to “read troves of documents to arrive at the paragraph that they’re really interested in.”
“So, we feel that generative AI has this complementary and copilot role to support our staff in expediting the decisioning, and bringing attention to what might be the real gist of what makes a good judgment call on a claim for example,” he said.
“We’re quite conscious about explainability and confirming how do we make that decision.
“Really for us at the moment, this technology is not deployed on a standalone basis. It is complementary and supports decisioning by humans, a staff member or other individuals in the organisation.”
Chauhan added the organisation is “very focused” on leveraging its partners and platforms” given the levels of investment they are making in understanding biases.
“This is where our partnership with our core platform providers make sense,” he said.
“We know we can’t invest to the tune on what Microsoft, and few other platform providers do, so we’re very invested in making sure we understand how they do it and then equivalently bring those practices into our workplace.”