Insurtech has both the data and the AI as it hopes to live up to the hype

Tired of hearing about glossy insurtechs that achieve very little before fizzling out?

Then Meet Tractable, a refreshing reminder that insurtechs can live up to the hype.

In under five years, the fast-growing artificial intelligence insurtech has grown from zero to 100 staff.

Tractable operates in eight countries – UK, France, Poland, Germany, Portugal, Belgium, Japan, North America – and always working with the top three players in motor.

As Insurance Times met co-founder and president Adrien Cohen at their trendy WeWork office in Old Street, Tractable is just about to launch in Japan.

Insurers just can’t get enough of them.

“We have just opened a very large office in Tokyo. Japan is one of the most exciting markets,” said Cohen enthusiastically.

“We have grown extremely fast, with very strong traction across the board because there is demand to deploy this technology and we are the only one to be able to make it happen.”

Image recognition firm

So what exactly is Tractable?

Tractable uses AI software to view images of vehicle damage and makes an assessment of the damage within seconds.

The technology connects with insurers’ back-office systems to help streamline claims management processes.

Alex dalyac

Tractable co-founder Alex Dalyac 

As an example, by having an image of the damaged vehicle, the insurer can instantly decide the level of repair servicing it needs, cutting down time and wasted money on initial inspections.

It might sound too good to be true, but in reality, all that has happened is that the AI has been fed thousands of accident-hit vehicle images and it can now instantly assess the likely damage.

The AI is self-learning, meaning the more images it gets, the better it becomes at assessment.

Ageas in the UK is a big user of Tractable. 

It is also very smart, which is why it took some very intelligent people to get the business off the ground.

That’s where the back story of Cohen and the two other founders come in, Alex Dalyac, chief executive, and Razvan Ranca, chief technology officer. 

Cohen is an ex-Goldman Sachs banker who left to help start up an e-commerce firm called Lazada following an approach from Rocket Internet, a German incubator that builds replicas of successful US tech business models in emerging markets. 

Lazada was eventually bought for $3bn by Alibaba in 2016, having become the largest e-commerce site in South East Asia.

During that time in Vietnam, Cohen had worked with machine learning enthusiast Dalyac.

Cohen, Daylac and Ranca, also a machine learning expert, teamed up to explore the ideas of an AI image recognition business.

They looked at pipeline corrosion, earthquake imagery and X-rays before eventually settling on damaged vehicles.

“We realised that this was the perfect market. There is a lot of friction in assessing damage to the car, it is a visual task – so the AI can learn it – and bodyshops have been taking pictures of damaged cars for decades.

“This is a goldmine for machine learning because you can use all those pictures of damaged cars to train an AI to understand damage to the car,” Cohen said.

Tractable angel funding 

To fund Tractable, they’ve used their own money, business angel capital and $35m of US venture capital.

Former Towergate chief executive Andy Homer and ex-Zurich claims head Tony Emms are backers who provided important insight.

“Those guys have been really helpful in understanding how to deliver value to the insurance company,” said Cohen.

Cohen said despite having all the foundations in place, it wasn’t easy at first convincing insurers to partner up.

“When we started a few years ago, we were the only ones. I remember going to insurance companies and they would really look at me.

“They were thinking ‘well, that sounds all great, but it is probably something in the future’.

“Now everybody is talking about that. Now our timing is better and that is why we are going to so fast.”

Tractable has both the AI and data, meaning it has a classic Warren Buffett ‘moat’ protecting it from being usurped by competitors such as Google, Microsoft and Amazon.

“The tech giants have the AI capabilities but they don’t’ have the data. We are the only ones with both,” said Cohen, who believes that by moving first they have gained a massive lead as the self-learning AI becomes more powerful by the day.

Looking forwards, Cohen wants to remain an independent business to exploit the limitless potential of AI. 

“Where AI is stronger than humans is that it’s much faster and it can operate at scale,” Cohen said, stressing there is a bright future for image recognition and Tractable if they ever decide to expand beyond damaged vehicles. 

We don’t doubt it. 

Subscribers read more

AXA plans to simplify e-trading for brokers


Axa suing genworth