EXCLUSIVE: “The Pursuit Of Accuracy” – Samuel Falmagne, Akur8 in ‘The Insurtech Magazine’

Akur8 Co-founder and CEO Samuel Falmagne is aiming to revolutionise actuarial science with AI that has nothing to hide

There is a mixture of optimism and caution about the role of AI in many industries, but one area that it seems destined to revolutionise is the actuarial function. After all, actuaries and AI share the same lifeblood: data.

Gathering it, analysing it, offering strategic recommendations based on it, and making the resulting insights clear and digestible is what actuaries do, day in, day out. Coupling their human expertise with increasingly powerful artificially intelligent tools potentially unlocks huge opportunities. And nowhere is there a bigger concentration of actuarial skill than in the insurance industry.

It’s here, where the financial cost of risk and uncertainty is constantly weighed in the balance, that Akur8 has focussed its ambition. Fresh from a funding round, and with a new foothold in the US, thanks to its acquisition of the award-winning P&C reserving software Arius, Akur8 has expanded to 160 people, with clients in more than 40 countries, since it was incorporated in 2019.

Co-founder and CEO Samuel Falmagne hadn’t come from an insurance background, but it was the opportunity for innovation in the sector that excited him. “I had a pretty long career in tech, at IBM, then I moved to an insurtech. I could see the potential of this industry in terms of disruption,” he says. “Insurance is a heavily challenged industry at the moment and needs to reinvent itself.”

That’s not lost on the big players, but it’s a colossal job, making it an appealing marketing for startups like Aku8 with targeted solutions. “When we started, a lot of domains within insurance were still untouched with very low technology adoption,” says Falmagne. “It’s also quite a stable industry economically that doesn’t suffer too much from macroeconomic changes – at least less than other industries – providing large, stable clients.”

Although entrenched attitudes in these legacy institutions may make them difficult to penetrate at times, there is a general acceptance that the prize is worth the pain. Investors certainly think so. A report from Gallagher Re found that funding in global insurtech saw a 39 per cent surge during Q2 of 2024 to reach £992million ($1.27billion).

It’s a refreshingly large jump after previous investment slumps.

“Machine learning and AI can only be used in insurance if there is a very high level of transparency around what’s produced by the algorithms”

“The risks in insurance are changing much faster than before,” explains Falmagne. “That is driven by climate change and by new technologies, such as electric and autonomous vehicles. You also have global inflation and other external drivers, which contribute to the pace of change.

“Clients’ needs are evolving, too, and people are behaving differently – COVID accelerated those changes. This pushes insurers to launch new products and evolve historical products.”

More innovative products and pricing models have a direct impact on the actuarial function. Take the case of a UK life insurer that now offers a product which monitors policyholders’ health and dietary habits, rewarding them with lower premiums with reduced underwriting. This gives policyholders broader coverage while making higher protection levels more affordable.

Not only do the figures show that clients stay healthier, but there is also a significant reduction in policy lapse rates. As data collection methods evolve to support these kinds of products, the pressure to adopt machine learning in the actuarial process increases. Falmagne points out that the insurance industry is in a privileged position when it comes to, not only how much data it gathers, but when it collects it.

“Even before someone becomes a client, they’ve already shared a lot of it with you,” he says. “That is very different from other industries. “If you look at retail, air travel, pharmaceuticals, you only start getting to know your client once they’ve signed up, perhaps by analysing their use of a loyalty card. Before that, you have relatively little information on them.

“In insurance, however, customers are asked a lot of questions during the onboarding process. There’s a huge opportunity to make better use of those significant volumes of data.”

Adding more human resources to process it is simply not financially viable – maybe not even physically possible. “AI and machine learning can be really transformational, to support the industry to make better use of all the data that it is collected, allowing insurers to react faster to market changes and client expectations,” says Falmagne.

“By applying machine learning to the process, the user can make better informed decisions.”

That last part is important.

Although AI’s impact on jobs is a legitimate concern here, as elsewhere in industry, Falmagne believes roles will not be lost so much as evolve in insurance. Analysts at EY tend to agree. They say there will be ‘growing demand for candidates with a blend of traditional actuarial skills and expertise in data analytics/AI’.

“I think the necessity for human involvement is one of the safeguards [against job losses],” says Falmagne, “but that means there needs to be some level of transparency in AI and machine learning technology.”

Indeed, he would go so far as to say that ‘machine learning and AI can only be used in insurance if there is a very high level of transparency around what’s produced by the algorithms’.

It’s important not only from an operational, but also a regulatory point of view.

“As soon as something has a black-box effect there is a risk that the regulator won’t accept you because, in each geography, there are pricing factors that you cannot use,” explains Falmagne.

Akur8’s solutions are based on this principle of full disclosure.

“We wanted to bring machine learning to the table, but keep full transparency and control over the models created,” he says. “Actuaries are manipulating a lot of data. By using machine learning in those use cases, we can supportnthem to do their job faster.

“We started with pricing use cases and recently expanded into covering reserving through our acquisition of Arius. We now leverage newer technologies such as software-as-a-service, the Cloud, and machine learning at each step of the process, whenever it makes sense to do so, ultimately allowing users to focus on decision-making.

“What started as one module has expanded to many but with the same vision of building a fully integrated actuarial platform and leading this domain within insurance.”

This article was published in The Insurtech Magazine Issue 11, Page 12-13