AI is a colleague, not a replacement – Aviva

Aviva’s global corporate and specialty (GCS) business does not believe that generative artificial intelligence (GAI) could supplant humans, but views the tool as a “colleague, not a replacement”.

The was according to Becky Brown, chief operating officer at Aviva GSC, who spoke during a panel entitled AI and machine learning in underwriting at the MoneyNext 2024 conference last week (20 June 2024).

GAI is a subset of AI that can generate media, such as text and images, with ChatGPT a prominent example of this technology.

Brown explained that GAI is considered a valuable tool within the business and aids in “micro efficiencies” such as “basic congestion and automation”.

She explained: “What I’m interested in is how far can we push it and that’s all in service of the underwriters using AI as a colleague – not to replace them.

“It is about sweating our assets – how can GAI get them to underwrite better and faster to meet the customer needs that we are trying to solve. And how do we reimagine the work in order to be doing that?”

In order to do this, Brown highlighted that the business was concentrating on two areas for using GAI in underwriting – reimagining wordings and controls to improve service efficiency   swift customer response and clear coverage understanding during issues.

She added: “Those are our two use cases where we’re currently reimagining what that roadmap could look like if we embed GAI.”

Regulation

Brown noted that the anticipated benefits are proving to be “interesting” as the insurer prepares to fully integrate AI.

She explained that the insurer is reviewing its wordings library to meet regulatory requirements.

This follows the implementation of the FCA’s Consumer Duty regulation on 31 July 2023, which introduced a comprehensive set of updated rules for firms.

These regulations have been designed to establish elevated standards of consumer protection within financial services, prioritising the needs of customers.

This included new requirements for firms to provide timely, clear and understandable information about products and services, so that customers can make good financial decisions.

Brown continued: “We’re examining our clauses and risk definitions and it’s revealing inconsistencies – like multiple definitions and interpretations for the same thing.

”Initially, we thought AI would speed up information retrieval for colleagues and underwriters but, unexpectedly, it’s also helping us tidy up these foundational issues.

“We’re also learning from our colleagues in other markets, despite differences in products and customer bases, to adopt best practices in governance and controls. The anticipated benefits are proving interesting as we prepare to fully integrate AI.”