Be’eri Mart: AI implementation can be managed with lessons from the past

As chief product officer at insurance software firm Earnix, Be’eri Mart faces the ongoing challenge of balancing a strong focus on the firm’s core product set – which includes rating, pricing and underwriting technology – with the constant pull of innovation, ensuring that technological progress does not come at the expense of stability.

In Mart’s words, Earnix aims to allow carriers to “define what their strategy is using analytics”, then “simulate the outcome of different decisions both in risk selection and pricing” and “quickly deploy that strategy into the market”.

And, given the technological and analytical nature of Earnix’s products, one key tool has become ever-more vital in accomplishing Mart’s day to day work – artificial intelligence (AI).

Speaking exclusively to Insurance Times, Mart explains that while today’s technological gold rush might seem entirely novel because of the cutting edge iterations of AI, lessons can still be learned from the past on how firms such as Earnix can differentiate from other digital driven prospectors.

He says: “We can use a comparison to something that happened a few years ago – the adoption of machine learning (ML).”

According to a 2022 article entitled A Brief History of AI in the P&C Insurance Industry, published by Loveland Innovations, ML first started being used commercially by financial services firms in the 1980s, primarily to assist with underwriting.

Mart continues: “There are many patterns we can learn from. For example, back then you had the introduction of ML platforms with lots of options, anyone could make their own model and had their own style. It’s the same with AI, anyone can make AI agents with ChatGPT, that’s not the uniqueness.

“What we did really well with ML – and we’re applying the same processes to AI – is we focused on the operationalisation. How do you not only put the right guard rails in place, but also have a process that allows business orientated people who aren’t tech savvy to use this in their workflows?

“And since we provide those workflows – that’s what they use today for pricing – it’s easy for us to help embed those agents into those workflows. That’s our differentiation.”

Balancing stability and innovation

Despite Earnix – and the insurance industry at large – working to elevate its offering in the recent AI era, Mart explains the importance of keeping the firm’s core product lines firmly at the centre of its focus.

“We recently conducted an internal analysis of what percentage of our engineering budget and headcount goes into our core product [versus AI initiatives] and it’s more than 60%,” says Mart.

“The core offering is not going away. If you want the workflows to operate as they are, then you’ve got to continue to invest. We’re not looking at [AI initiatives] as a cost reduction, we’re continuing to grow our employees – but now we’re doing much more.”

Mart adds that Earnix is also using AI internally, for example helping developers to generate code or tests. For these use cases, however, he does not think the technology is currently mature enough to work without human oversight.

Looking forward though, Mart predicts that it is inevitable that AI will become the “owner” of insurance workflows, at least in certain domains.

He says: “I absolutely see a world [in the future] where certain things that are done today by humans will not involve humans at all. I absolutely believe that.

“We should not be afraid of that, we’ve just got to figure out how to get ready for it. But at the same time, I think the big decisions where there is reputational and brand risk, customers will make that the last phase of taking out humans.”

Mart notes that AI is producing an entirely new suite of roles too – ones that appeal to a younger, tech native generation and may offer an alternate route to filling industry-wide talent gaps.

He continues: “There is a problem with new hires. We actually call it the ‘junior crisis’ because sometimes it’s not easy for them to find a job.

“But this is the perfect job for them. We can get the best talent. They’re really smart people, they’re born with the world of AI and they can do phenomenal things. We have a whole programme around hiring those kinds of people.”

Accelerating adoption

In April 2025, Earnix announced its purchase of insurtech and generative AI platform Zelros. This move aimed to accelerate the firm’s AI capabilities and add an extra dimension to its product offering.

Mart says: “From a business perspective, Earnix’s focus has traditionally been on the strategy and operations side, where the pricing and underwriting teams are. But we always depended on another platform to help operationalise it in the field.

“Zelros, because [it comes] from the side of digital channels, [its] focus has always been on delivering [AI tools such as next best actions, magic answers, automations and call summaries].

“The thought process was ‘let’s bring the two together’ and help become the entire cycle and make the entire cycle more agile.”