There is still ’an enormous amount of work to do’ around introducing generative AI, says group chief underwriting officer

Insurance firms are ”quite optimistic” about their time frame for implementing generative artificial intelligence (AI), but having patience around project deliveries will be essential for success, according to Miqdaad Versi partner at Oxbow Partners.

Versi was speaking at the launch event of Oxbow Partners’ The maturity of generative AI in the specialty and reinsurance markets report, which was published on 2 July 2024.

The report was based on interviews with senior leadership at 22 of the largest global reinsurers, insurers and specialty firms. It was conducted between April and May 2024. 

Addressing event attendees at Lloyd’s of London on 2 July, Versi said: “Over 60% [of report respondents] thought that the industry would be in the strategic foundation phase in 12 months [time], which is quite optimistic”.

Highlighting some of the report’s key findings, Versi noted that 19 respondents believed generative AI would be transformational over the next five years.

Eight insurers said generative AI would be “entirely transformational”, while only three said the technology was “not at all transformational”. 

Furthermore, 15 insurer respondents are currently in “tight experimentation” with generative AI projects, while 62% of insurers already have an internal chatbot or one in development.

Versi continued: “It takes time for that innovative culture to be built, people to understand how they could reimagine their processes and to understand how [to] change [their] business for the better.

”The only way that can happen is by a strategic decision to push this forward from the top, technology and data foundations are required to enable [it] to work.”

James Slaughter, group chief underwriting officer at Apollo, added: “It’s quite hard to scale [generative AI] – you need patience.

“It all starts with the basics of having the right technology in place in our industry. Principally, we’ve got an enormous amount of work to do.”

Despite agreeing with Versi that patience is required and that the insurance sector shouldn’t rush into AI implementation, he also noted that ”we as an industry have a horrible habit of taking too long to get stuff done”.


Expert panellists attending Oxbow Partners’ event additionally flagged the importance of data, particularly in terms of fully utilising generative AI.

Slaughter continued: “Data is clearly a competitive advantage for us as an industry. We’ve got more comfortable sharing data more broadly. If you were to speak to very senior people across the industry, there’s still a tendency to say ‘my data, my insights’.

“Until we create huge accessible datasets, these tools are way too sophisticated for the relatively small data we have.”

For Slaughter, patience is also needed around building data and talking to each other “in a way that allows us to use that data”.

Marianne Harvey, chief operating officer at Aegis, said that data must be “especially clean” for use with generative AI.

Cost and education barriers

For Kanika Chaganty, chief data officer at Brit Insurance, “education is really important” when it comes to AI.

Chaganty explained that there has to be understanding across organisations around how to use data and technology in a way that aligns with the vision for the business.

She continued: “The technology is evolving, people have to be very patient. We have seen in the early days of experimentation in the past year, depending on what question you ask and how you ask [it], you will get a different answer.

”You have to have the patience to invest the time to train these people, you also have to be patient around the bandwidth because it won’t [happen] overnight.”

Slaughter added: “We don’t expect innovation to throw [return on investment] in the first 18 months. It comes back to patience – it takes a long time to demonstrate, certainly on financial metrics.”