Health and life insurance providers face a rising volume of claims due to chronic disease and aging populations, but their processes lack scalability. AI automation can help, but it takes more tailoring than for other insurance types.
Automating property and casualty insurance, for instance, gave rise to public company Guidewire decades ago, but health insurance is a tougher vertical. “You can’t automate in a simplistic way, like you could for a broken car windshield,” Paris-based entrepreneur Tarik Dadi told TechCrunch.
His startup, Qantev, is hoping to solve this. It provides clients like AXA and Generali with software that helps them manage claims via AI models that go through the same checks as in-house medical staff currently do: “Is the care medically necessary? Is the price right? Is the bill fraudulent?” But it does it much faster, which helps reduce costs and customer churn.
Dadi saw that need while working as a senior data scientist at AXA, while now-CTO Hadrien de March, a doctorate and former quant, had the math chops to address it. The two of them joined forces at Entrepreneur First in late 2018. “EF’s concept is ‘pre-idea, pre-team,’ but I cheated and brought the idea,” Dadi said.
Armed with an idea and a small team, Qantev went on to raise a €1.7 million seed round led by Elaia in 2020, followed by a €10 million Series A round led by Omnes and Raise Ventures in 2022. These three VC firms are now participating in Qantev’s €30 million Series B round, which happened sooner than planned, Dadi said.
“Our topics are quite hot at the moment, and we saw that YC, at the beginning of the year, included at least three of our topics in their wishlist,” he said, referring to Y Combinator’s Request for Startups and to what he calls the “LLM craze.” “We started to see lots of small startups popping up in the U.S. and just throwing an LLM at the problem. […] We know that it’s a hard problem and that we have an asset.”
One conviction Qantev developed over the last five years is that one large model is not enough; its software relies on a collection of AI models trained on historical data from its clients and aiming for accuracy. “You can’t have hallucinations or anything like that. It’s human health; you can’t refuse care for someone’s cancer. That’s why we are still a big AI shop. We have many PhD and ML experts in our team because we have to create small AI models that are highly specialized in our topics,” Dadi said.
Qantev is aware it could still get leapfrogged by newcomers, and the company plans to use its new funding to recruit the AI and engineering talent it needs to maintain a technical advantage. Its goal is to double its headcount by the end of the year.
Led by Blossom Capital, the Series B round will also support Qantev’s international expansion; it plans to grow its Asia-focused Hong Kong office and make a strong push in North America.
While it has competitors there, such as Alaffia Health and Anomaly, other Blossom portfolio companies have made strong headways in the U.S. in recent years, and Qantev has an advantage of its own: Its customers are large and global, generating organic expansion when a new subsidiary adopts its software.
The downside of targeting such large clients is that sales cycles are long and complicated. “But the upside is that they’re big-ticket items,” Dadi said. He liked Blossom’s understanding of enterprise software as a category, and of Qantev’s ambition to become an operating system for health insurance. “We like saying it is a platform, because it’s going to be multiple products.”
What these products might be remains to be confirmed, although underwriting seems to be a strong candidate. For now, Qantev prioritizes claims management, but it is easy to see how it could leverage the legitimacy and data access it is gaining from its early customers to help them streamline other operations, as it is already doing with fraud detection. Looking at the big picture, this would tie back to the trend of AI as a way to fight rising healthcare costs.
You must be logged in to post a comment Login