Insurance has been fertile ground for artificial intelligence innovation, working as it does at the nexus of giant datasets, risk assessment, predictive analytics, fintech and customer service. Federato, a startup riding that momentum, has now raised $40 million to expand its business: an AI-powered underwriting platform to help insurers better understand and respond to risk.
Stepstone Group is leading the round with previous backers Emergence Capital, Caffeinated Capital and Pear VC participating. The startup has now raised $80 million in total. It is not disclosing valuation, CEO and co-founder Will Ross confirmed in an interview that it was a “serious, significant upround” that was multiples bigger than its previous valuation.
For some context, Federato’s last valuation was $125 million based on a fundraise from last year. Further context is that Duck Creek, one of its competitors, was snapped up by Visa Equity for $2.6 billion in 2023. The latter company offers a wider set of SaaS for insurance companies, it does show you directionally where the valuation for a lucrative AI product aimed at this sector could go. (Insurance, globally, is estimated to be one of the biggest industries in the world, totalling multiple trillions of dollars in value, with underwriting one of key areas where AI is expected to play a major role.)
Federato was co-founded by Will Ross (CEO) and William Steenbergen (CTO). Ross was one of the early employees at the Watson group at IBM, where he worked on the deal to buy the Weather Company and leverage the data from there to build environmental models. Later he went back to graduate school at Stanford, where he met the other William (Steenbergen).
It was 2021 and AI was already all the rage (January 2021 was the month OpenAI released Dall-E, its image generator). But it was already being positioned by many as a replacement for repetitive work.
“We had a shared thesis that AI could better be applied to optimizing what no human could do or had time to do, versus automating away low value tasks,” Ross said. “The analogy here is Uber, DoorDash. These are consumer companies, but they solve a problem no human had the time to effectively solve. And those tend to look like these optimization problems.”
Ross was in the business school, and Steenbergen was in working in computational and mathematical engineering. “We were actually collaborating on a very different project, a wildfire modeling project,” Ross continued. “And we started to get to know the insurance industry, because they were interested in the wildfire modeling.”
That in turn led the pair to thinking about the wider challenges in insurance: typically when insurance companies are working through an underwriting process and coming up with how to cost a product against the risk around it; or even what products to build at all, there will be a team in-house bringing together vast amounts of data both directly and indirectly related to potential insurance scenarios, crunching numbers to get better insight and make more informed selections. Federato’s solution, which the company calls RiskOps, provides decision support around that process.
The startup claims that customers see 90% improvement in “time to quote” (how long it takes to give a quote on a particular service in aid of winning a sale) among other efficiency improvements.
The pair’s initial foray into modelling wildfires led to one of its early customers, Kettle: a reinsurance platform that has up to now mostly focused on one market, California, and one big problem: reinsurance and fire catastrophes, a big issue in the state. It also counts larger companies like Nationwide among its customers.
While speed is definitely something that AI brings to bear in its work, another important aspect here is how it is used to evaluate large amounts of data, an important part of the diligence that insurance companies must go through in their work.
“We are in the land of Facebook, and moving fast and breaking things,” Lotti Siniscalco, a partner at Emergence, said. “But in insurance you cannot do that.”
