’Generative AI [is] able to get all that unstructured data and make structured elements – [it] fuels the ability to create models,’ says senior vice-president 

Predictive models built using generative artificial intelligence (AI) form the backbone of market “overperformers”, according to Leo Tenenblat, senior vice-president and general manager of data and analytics at software company Guidewire.

Tenenblat shared this insight with trade press during a breakfast briefing on 19 November 2024, held during his firm’s Guidewire Connections 2024 conference in Nashville.

At the event, he explained that both economic and climate “volatility” is ”really widening the gap between underperformers and overperformers” in the insurance sector, with businesses that can access predictive modelling based on timely and high quality data often outperforming their market peers.

In turn, firms that are unable to adapt to fluctuating market conditions promptly – maybe through the lack of accessing good quality predictive modelling tools – risk being “adverse selected out of the market”.

Underpinning successful predictive models is the use of generative AI, Tenenblat continued. Generative AI – which Tenenblat described as being one of the most exciting developments with data in the insurance market today – refers to deep learning models that can generate high quality text, images and other content.

“Generative AI [is] able to get all that unstructured data and make structured elements – [it] fuels the ability to create models,” he explained.

Although Tenenblat is a fan of what “super critical” predictive modelling can bring to the table, he admitted that it is ”very difficult in many cases to get those insights embedded in the screens of underwriters or claims adjusters”.

He added: “We and other vendors are trying to standardise that, to make it a lot easier and to bring in the explainability for those insights.”

Critical models

When looking at modelling in general – whether a predictive tool or a pre-built model that is ready to use – Tenenblat pinpointed pricing as ”probably one of the most important models that people are building” today because ”the importance of taking into account volatility is very critical”.

Guidewire has been building tools to make it easier for carriers to build such models. 

Will McAllister, managing director of Europe, Middle East and Africa (Emea) at Guidewire, said: “Historically, bringing products to market has been a challenge for insurers because of legacy systems. Once you get your product to market, you need to be able to rapidly adjust that product.

“If you can’t adjust to changes [in] the macroenvironment, you are going to end up losing money. It’s that agility and flexibility that insurers are going to need moving forward. 

”For us, that’s why a fully open platform that enables that ingestion of data, enables the transverse of data across the [proposition] to get to [the] point of impact is so important.”