As a software engineer with Google, Tomas Vykruta (pictured) served a stint at what became Waymo – the technology giant’s autonomous driving tech development company. He was tasked with helping its test car work better after a widely publicized crash into a bus.
A rebuilt system based on machine learning was the answer, and the vehicle subsequently worked much more intelligently “with an infinite number of rules” driving its functionality. After the experience, Vykruta began to increasingly wonder what other industries that hadn’t modernized yet could benefit from the technology. Insurance came to mind, after which he and other co-founders came up with EvolutionIQ.
The insurtech start-up relies on a machine-learning “claims guidance” platform that helps insurers dealing with bodily injury claims organize their time and focus on the most actionable claims, with a goal of achieving the best outcome for claimants and carriers alike.
Insurance, particularly the claims space, was an important target for EvolutionIQ, Vykruta said.
“Insurance is something that we all end up paying for,” he said. “If we can reduce the cost of insurance, it’s amazing. People can spend more money on things like putting their kids in better schools, going on nicer vacations – it just increases the standard of living for everybody.”
The company noted in April that it had raised $21 million in Series A funding. Brewer Lane Ventures led the round with participation from previous investors and other strategic investors including Reliance Standard Life, New York Life Ventures, Guardian Life and Sedgewick, a global risk and benefits company.
EvolutionIQ works with chief claims officers at disability and P/C carriers as well as TPAs (Third-Party Administrators) looking to improve loss and expense ratios as well as customer services. Customers include Reliance, Sunlife, Reliance Standard, Principal, Argo and Matrix Absence Management.
Vykruta claims the 25-employee company has created novel AI-driven claims guidance technology – a platform that works as a co-pilot for claims operations.
“We’ve deployed this with a number of the top 10 and top 20 carriers in disability, and there’s nobody that’s doing this today,” Vykruta said. “It’s a problem they’ve been looking to solve for a long time and the technology just wasn’t available.”
The reason this issue hasn’t been tackled so far, he said, is because of the complexities of bodily injury claims.
“It’s not like a transaction on an ATM card. It’s more like a book with many chapters and requires a human expert like a doctor for a read through for the whole thing, to be able to understand what’s happened and make a forecast of what will happen next,” Vykruta said. “There are thousands of ICD [medical condition] codes … and a single claim can have as many as 40 [ICD codes] overlapping over a number of years. So you can imagine it will take hours just reading through a single claim, just to get caught up on where it’s at and then you have to make sense of all this medical complexity.”
EvolutionIQ has made the job much easier, he said.
“Our system has medical [expertise] and it reads through the entire claim, and this is only possible with deep learning, which is a technology that really only became mainstream in the last couple of years,” Vykruta said.
Insurers also have increasingly digitized their data over the past 20 years, and that has enabled a platform like EvolutionIQ to work effectively, he added.
Deep learning, a form of machine learning, is a key part of the platform, as is recurrent neural networks, a variation of deep learning that can reason about a sequence of events, such as what happens in the weeks and months after someone is injured or becomes ill.
Data ingestion technology is another key ingredient - it helps absorb vast amounts of data, both low and high tech.
“We take all of the carrier’s historical and active claims data, including all of their unstructured, hand-type notes and attachments, and that all gets sent through a neural net,” Vykruta said. “Every day, our system sees all of their active claims data going back years.”
The usual option would be to hire an auditor to go through what can be complex and lengthy claims files. EvolutionIQ does this every day, the company’s founder explained.
“We go through every single document, every note, every claim, so nothing is left unturned,” Vykruta said.
The platform also relies on what Vykruta describes as a “medical explanation system,” which is technology constructed to be able to understand nuances of various medical diagnoses.
“Our technology understands how they interact with each other. You know what the duration of the outcome is, and what the outcome is for recovery,” he said.
Aiding the process, the platform links a carriers’ data set with “very large medical dictionaries and libraries” that have vast amounts of data about historical recovery events, according to Vykruta.
The user interface is also designed to be elegant and easy to learn, he noted.
EvolutionIQ is in the bodily injury insurance space, which can include disability cover, workers’ compensation and general liability insurance.
Find out how much does general liability insurance cost here.
“Convincing the carriers initially to work with us was really hard. We were a group of really smart AI engineers, but we hadn’t had a lot of experience with insurance, so we had to convince them that we have a new way of looking at things and we can solve this,” Vykruta said. “Once the carrier can see just how well this technology works and what it is doing with their book of business, they are then rapidly asking us to solve all kinds of other problems for them.”
