Optimizing your Insurtech for venture investment

LearningsApr 3, 2024

Written by Dave Wechsler.

The Insurtech category is starting to bounce back. Coming off investment lows of ’22 and ’23, there’s renewed interest from the VC community in Insurtechs and MGAs (managing general agents). But the bar has been raised considerably. As an investor, I’m often asked what it takes to get funding in today’s market. This is the first installment of a 3-part series meant to provide some insight into what OMERS Ventures found attractive about each of our insurance investments: Clearcover, Foresight and Joyn Insurance. First up — Clearcover.

Clearcover — a story of data differentiation
You may have seen some recent press on an OMERS Ventures portfolio company, Clearcover, and the launch of its latest AI product, complementing its ClearAI ® to expedite the claims process. If you aren’t familiar with Clearcover, it’s an Insurtech OG — founded in 2016 with the vision that a digitally-native personal auto carrier could offer a superior product at scale. They knew one key thing to be true: When optimizing insurance products, the more data, the better. Unfortunately, legacy insurance systems store data in siloed, unstructured states, making it unavailable to the rest of the enterprise. So Clearcover built the technology themselves to make that data accessible.

Clearcover architected its foundational systems around the notion that every piece of data collected offered unique insights. Whether generated in underwriting, collected during claims, or aggregated by a third party, a core tenant of the Clearcover tech stack design was prioritizing data integrity and availability. Also key to the Clearcover vision was its “API-first digital substrate.” Essentially, that means every piece of data is available via a simple API call. This tech-forward approach was expensive to build, but novel for an insurance provider, and core to why we invested.

And the bet paid off: The initial sales delivered data the likes of which the team had never seen before. As the data set grew, so did excitement around leveraging machine learning (aka “AI”). The first obvious AI use case for Clearcover became apparent in 2020. By this time, the team had tens of thousands of quotes and policies to work with, and the engineering team developed an ML retention model for decisioning within their acquisition funnel. The accuracy and success of this initial model led to a burst of new ideas at the company, all of which were soon bundled into the launch of a more comprehensive AI offering called “ClearAI.”

ClearAI has had a direct impact at all levels of the organization, but the company is currently leaning into two high-impact areas to maximize results: marketing optimization and claims experience.

Marketing optimization
Insurance is an industry built on trust. Customers who are happy with their insurers act more like a partner, staying longer with the company. The happiest Clearcover customers embrace a digital-first experience. And while there are demographic indicators of who these customers are, it’s the psychographic characteristics that seem to best indicate a strong match for Clearcover.

As ClearAI analyzed hundreds of thousands of quotes and customers, their machine learning reversed-engineered the pipeline to see what marketing tactics and value propositions most effectively attracted those ideal customers. They struck gold: Eventually, the AI model advanced to a point where it could handle real-time channel optimization. So now, Clearcover is empowered to increase acquisition spending with confidence, and as the book’s KPIs change, the sales paths respond to optimize new business close rates.

The claims experience
ClearAI’s second big tent pole is around claims. While the most critical aspect of the customer relationship, the claims experience isn’t always positive. In fact, the simple eligibility of a claim is still a point of friction with incumbents, who often take days or more just to confirm or deny coverage. With ClearAI, many claims can be confirmed eligible within minutes of filing (FNOL), giving customers peace of mind when they need it. Faster claims response times help Clearcover get ahead of loss adjustment expense (LAE) and negative loss development.

The ClearAI FNOL process is remarkably straightforward: The insured party can easily start a claim in the Clearcover app. The customer will answer simple questions and easily submit photos. ClearAI then determines if the claim is eligible for Clear ClaimsTM — a product that instantly determines coverage eligibility and allows certain eligible claims to get paid in under 30 minutes (their record is 7 minutes!). That’s not all ClearAI does during FNOL. However, the insured party is then offered the chance to interact with an AI assistant. Unlike some of the more erratic popular general-purpose chat-bots out there (looking at you Grok), the AI assistant asks friendly questions and assumes the insured is likely in a distressed state. As the AI agent analyzes feedback, it automatically fills in necessary data points it extracts from the conversation. Once all data points are collected, the conversation ends, and the customer can then review the answers the AI pre-filled and confirm their accuracy before submitting. Over 50% of claims filed now use this channel.

This simplified experience has proven to have a positive impact on customer satisfaction scores. ClearAI turned what is typically the worst experience with a carrier, into one that actually affirms a customer’s insurer has their back.

Data is the differentiator
While early investors in Clearcover knew it would take time to build these data sets, ultimately the bet was on the company’s ability to leverage the data in ways that legacy carriers could not. For a digital-first company like Clearcover, which has 6+ years of detailed customer data, that adds up to a differentiator that few can challenge. This puts the company in a position to quickly test and deploy new data-hungry models. And as LLMs become more powerful, the ClearAI team will leverage this rich dataset to train new models to further improve their operations and customer experience. Incumbents, meanwhile, are forced to focus their efforts on maintaining, and ultimately migrating, legacy technology.