Free Preview

This is a preview of some of our exclusive, member only content. If you enjoy this article, please consider becoming a member.

InsurTech Coterie Insurance is on a mission to bring speed, simplicity and service to small business insurance—and it’s accomplishing that with a 100 percent digital underwriting process.

Does that mean the managing general underwriter doesn’t actually employ insurance underwriters?

In fact, according to Co-Founder and Chief Executive Officer David McFarland, underwriters play a special role in the company, where they are engaged in activities vastly different from the ones that consumed their days at traditional insurance companies where they got their hands-on experience as underwriters.

“Even our head of insurance comes from the underwriting discipline,” reported McFarland, who had a decade of experience as a casualty actuary before launching the InsurTech. “What our underwriters are really focused on is programming their brains into rule-based engines,” he said.

“We’re working with fairly large amounts of data and risks that can be understood at a high level based on the data. We’re working with smaller and micro commercial businesses. This is not large specialty operations. So, when we look at these risks, we can take what we know from an underwriting discipline and use technology and data to programmatically decide what risks we don’t want, what risks we do want, how we should attach endorsements, how we should utilize IRPM [individual risk premium modification] schedule credits, debits, etc., in addition to our rating algorithm.”

“Over time, as we see these effects take place, we see how the book performs, we can learn even more and create more rules—both in terms of controlling cat exposure and in terms of ensuring that we get the risks that we want and avoid the risks that we don’t, or provide the right terms to make it the risks that we do want. Sometimes it’s not necessarily [that] an entire risk is bad, but we’ll agree to insure the risk based on certain terms.”

“We use our underwriters to inform how we do that since they have expertise in the space,” he concluded.

In other words, Coterie Insurance’s underwriters are tasked with big-picture underwriting—putting all their knowledge into rules rather than looking at individual risks one by one to decide which to accept. There aren’t any underwriters working at Coterie Insurance that are directly dealing with an agent or customer and declining or accepting a risk.

“That’s exactly right,” McFarland agrees. “Think about it this way. If you were the world’s best underwriter, and you were mainly looking at risks that were not super-complicated specialty risks or anything like that, but you were looking at these smaller insureds, [then] if you could clone yourself, teach yourself all the rules, I think most people would legitimately do that, especially if you remove the more subjective parts of your job from it, like the dealing with the agents,” he said.

“What we’re doing is we’re just making these excellent underwriters more effective by enabling them, with technology, to do their jobs a hundred thousand X what they would be able to do otherwise.”

“We have a pretty strong insurance background. We believe this is necessary,” he said, when asked about the makeup of the Coterie Insurance staff. “If you want to help push the insurance space forward, you need to couple that with not just falling into the same line of thinking that most others would,” he said, referring to the rest of the team, with expertise in engineering, data science, UX design and other non-insurance areas.

What about actuaries? Carrier Management asked McFarland, a Fellow of the Casualty Actuarial Society, who started his actuarial career with the National Council on Compensation Insurance, what actuaries do at Coterie.

Like the underwriters, he said Coterie Insurance isn’t employing actuaries in a traditional way. “It’s more like we employ their talents and their knowledge to the end of structuring and building out things in a way that we can multiply their effect. And I think that’s kind of the general gist of what I’m trying to say across underwriters, actuaries, etc. If you just think, ‘How do I 10X this? How do I 10X this without putting any more money behind it, right?’ If you just force yourself to think about those things, you can really stretch your mind and do some pretty incredible things.

“So, how do I 10X quote unquote the effectiveness of an actuary? How can I make it so that one underwriter can manage a book that’s a hundred million in premium or more?”

“Those are the things that we’re challenging ourselves with here at Coterie. And so, when I talk to an actuary, and we look at [the question] how do 10X their effectiveness, we can utilize things like machine learning. We can utilize different data science models to really understand the book in novel and efficient ways—and apply those learnings programmatically and almost automatically.”

“What we’re doing is we’re just making these excellent underwriters more effective by enabling them, with technology, to do their jobs a hundred thousand X what they would be able to do otherwise.”

David McFarland

“Think about it this way: One of the most inefficient and illogical things that we do on pricing, but it falls [to] almost every actuary, is [calculating] frequency and severity trends. Goodness gracious, that stuff has so much of an impact on rate, [but] it is just nonsense most of the time.”

“What happens is you’re going to pick this X-point trend, and then the product manager is going to want this [lower] trend because his agents are pushing him in this direction, even though that results in a negative rate increase or a rate decrease, and you really need a 4-point rate increase to stay actuarially adequate within the region.”

What the exercise amounts to in McFarland’s mind is “poorly done pricing optimization that’s just wrought with bias.”

At Coterie, “what we do is we say let’s not use these techniques that are just filled with bias, and instead, let’s provide actuarially adequate rates based on what we’re seeing in actual data and true trend analysis. Those are the things that we employ,” he asserted.

Data Sleuths and Two-Question Binding

While Coterie Insurance utilizes the inputs of data providers that crawl the Web for information about small businesses and uses AI to validate the data and provide risk insights to carriers and MGUs including Planck—profiled in a separate article that is part of this series on underwriting—Coterie Insurance has built its own decision engines to make sense of all the incoming data. (Related article: “Planck: Mining Data Gold for Small Commercial Insurers“)

“It’s a mistake to think you can just hook up to anybody and then you can have automated underwriting. It’s not really a thing. You get the data in, then you have got to make decisions on it…Then you have to operationalize that,” McFarland said, noting that Coterie Insurance built technology that it refers to as Sleuth—”because we’re nerdy actuarial and underwriting folks”—to pull in all the data “and run it through imputation and machine learning models to figure out if the data is reasonable.”

“How much confidence do we have in this stuff? What can we impute based on these statistics? And then, subsequently, how do we now use this information? Are we going to endorse this policy? Are we going to just exclude it? Are we only going to exclude it on a property side? Are we going to exclude it on a liability side?…How can we make this risk desirable?”

“It goes through our whole engine, and part of that engine is fueled by underwriters,” he said, explaining that Sleuth has a user interface into which “our underwriters can, with their keyboards,” create rules from 3,000 incoming data elements. He shared a hypothetical example: “If this particular field is greater than X and it’s in Florida, then we want to endorse this wind-hail exclusion.”

“They can push that to production completely on their own without a developer or anything like that. Our team did this so that we can automate the operational side of underwriting to where we’re not having to wait multiple weeks to get a new rule in. And that’s pretty magical. We’re really happy about the effectiveness that our underwriters can have to take moratoriums, [for example], to really take a scalpel to particular risks and just help us build a book dynamically and intelligently.”

Coterie Insurance distributes small business insurance through agents and brokers and through partnerships with financial software providers, PEOs and other business platforms. “We don’t do any D2C [direct-to-consumer]. And even the partners who are more non-traditional still have their agency licenses,” McFarland reported.

In July, the InsurTech announced an iteration of a dashboard for agents and brokers known as SimplyBind. The big news: SimplyBind now prefills information based on two data points: just business name and address.

“Using Sleuth, we get all of the data” and return to the agents with a decision: “‘We want to write this risk and we’ll do it at this price’ or you’re going to get a declination with a reason,” McFarland said. “And that’s going to be in seconds, by the way.”

“What this ultimately does is it turns the 32 million small businesses out there, which were frankly more of an obligation to agents and brokers before, now they’re an opportunity. Now they’re actually something that they can write efficiently and profitably,” he said. “This comes as a relief, I think, to most agents and brokers because these small commercial businesses are growing. They need coverage. And these agents and brokers need to be able to write the business profitably. We enable that.”

“Our whole focus is bringing speed, simplicity and service to the commercial insurance space. And when we started with these agents and brokers, I believe our average bind time was 7.9 minutes. And now with SimplyBind and Sleuth, people are binding policies in 3.4 minutes,” he said, stressing that time frames he cited were averages, not minimums. “You can go much faster than that if you’d like.”

SimplyBind is available for small business owners package policies, general liability and miscellaneous professional liability, according to McFarland. “We did launch workers comp, but it’s on a very small scale. We’re just testing the pipes right now. We’re going to really expand that out to our partners in 2023.”

One of the intriguing changes that came with the July rollout of SimplyBind, compared to earlier iterations of the platform, is the fact that Coterie Insurance is not asking about claims experience of potential small business customers. Agreeing that past claims experience is “definitely predictive” of future losses, particular for larger risks, McFarland explains that doesn’t mean Coterie Insurance has to know exactly what claims happened when, instead relying on connected insights from the 3,000 pieces of data it collects for each business name and address.

“We think that there are some things that are highly correlated with prior claims experience that we are able to leverage…If I Google reviews of a few [restaurants], and I see comments like, ‘Every time I come into this place, the bathrooms are filthy and water’s leaking out the faucet,’ [then] that sounds like a bodily injury liability claim waiting to happen.”

“There are all of these data points, [and] even though those are readily available to us, [the industry] has been neglecting that and saying, ‘I’m sorry, the only thing that I care about is if this person had a claim.’ However, that is not the most indicative of future claims experience. I actually think that these other data points can be more predictive if they’re used intelligently,” McFarland said.

In addition to Planck, McFarland said Coterie Insurance uses a lot of data vendors, declining to disclose how many. “No single data vendor has all the answers. We’re huge fans of Planck, but to only use one data vendor, you may be missing out on some things that are very pertinent to the risks that you write.”

“You may be able to get away with just one data vendor depending on what you write. But for us, we’re doing BOP, which is both liability and property coverage. We’re doing professional liability. [So,] I think every data vendor out there.” One vendor may deliver roof scores for property risks, for example, while others may not. “When we look at these types of things across the variety of vendors, we pick out who’s the best in which particular area. And we do have redundancy even—to make sure that we feel good about the risks that we’re writing and the data that we’re collecting on those risks.”

Just because you have data on something doesn’t necessarily mean it’s right. So, we want to make sure that we have the necessary data from all of these things to make sure that we can do this intelligently and it’s beneficial to the end consumer and the agent.

Asked whether coverage or explainability or customization were important in Coterie’s selection of data vendors, McFarland said that “it’s API response time because we do everything via API…So, make sure that you can send us this fast. API response time is so crucial for us because we’re trying to create a delightful end-user experience for our agents and brokers and policyholders.”

Accuracy is another consideration. “This is another reason why you have underwriters. You go through with your underwriters. You figure out what are the actual exposures, classifications, etc., for these risks, and you match them up against the data providers. And who’s going to come out on top? So, speed, accuracy, pricing…And, of course, is it relevant to your book of business?”

This online bonus article complements features that are published in Carrier Management’s fourth-quarter magazine about how carriers and MGUs are transforming commercial underwriting using AI and automation with the help of technology partners.

Related magazine articles include:

All of the articles in the magazine are available on the magazine page of our website.

Click the “Download Magazine” button for a free PDF of the entire magazine.

To be able to read and share individual articles more easily, consider becoming a Carrier Management member to unlock everything.

For Coterie, the bells and whistles that data providers may add to their offerings, such as continual underwriting or seeking additional risk factors, aren’t a big draw. “Honestly, we prefer just the raw data. We don’t even want the insight. Just give us the raw data that you scrape, and we’ll make sense of it. We have a pretty high-powered data team behind us.”

Mission Accomplished?

For Carrier Management’s feature on commercial insurance underwriting transformations, we also interviewed QBE North America executives, whose mission is to simplify the day in the life of an underwriter, and Nationwide, where the mission is to be relevant to distribution partners and customers. To conclude our interview, Carrier Management asked McFarland to restate the company’s mission—and the problems it’s trying to solve.

“Our mission is to bring speed, simplicity and service to commercial insurance,” McFarland said. “This is going to create two things, we think. One, it’s going to really benefit the insurance distributors…Before we came along, in this whole space, you had these 32 million small businesses. Most of them are uninsured because agents, brokers are losing money whenever they touch them. So, we’re benefiting them.”

“We’re also benefiting us as an insurance entity, right? We’re lowering the expense ratio because we’re not touching [the business] all the time…And with our additional data, we are getting insights that others aren’t. So, we’re actually able to improve the loss ratio.”

“We’re tackling both the expense ratio and the loss ratio—really focusing on the holistic view of the business in order to improve things overall,” he concluded.