Digital Whitepaper · Personal Auto
The author
Three decades of leadership across virtually every discipline of an insurance company — and an actuary's eye for what actually moves the combined ratio.
Executive summary
Since credit scoring in the mid-90s, telematics in 1998, there have been no transformational variables introduced into personal lines insurance. After 25 years, that drought is over. Insurers now have a novel, ground-breaking way to determine who, or what, to trust.
Economic benefit of using trustworthiness as a variable — personal auto alone.
The cost today, paid by insurers and the trustworthy majority in friction and missed fraud.
New business, renewals, claims and billing — one variable, applied at every touchpoint.
Bad actors introduce inaccurate information and fraud into the system. Insurers respond by adding friction and process to catch them. The genuine majority pay for it through higher-friction journeys and higher premiums. With Clearspeed, insurers can now assess the trustworthiness of applicants, claimants, and agents in seconds, at any point in the policyholder lifecycle. This commentary lays out the impact Clearspeed can have across auto insurance lines.
01 — The macro shift
Trustworthy systems lower costs and friction. But a quarter-century of grievances is eroding trust broadly. The days of the roadside farm stand running on the honor system are disappearing, and pharmacies now lock up everyday necessities, treating every customer as a potential bad actor.
Interwoven with declining trust is the rise of soft fraud. And consumers themselves are justifying exaggerated claims because they mistrust insurers.
of Americans believe it's acceptable to inflate an auto claim under certain circumstances. (CAIF / Verisk, 2022)
might pad a claim if they felt their insurer wronged them or was being difficult. (LendingTree)
European consumers would consider exaggerating a claim if treated unfairly. (PwC, 2022)
The industry's response has been understandable: add process and friction, accept that some fraud won't be caught, and raise rates for everyone to cover both. Bit by bit, the US insurance system has evolved into a non-trustworthy system, with the genuine majority bearing the cost.
02 — The cost of non-trustworthiness
Using personal auto to illustrate, we size the cost of non-trustworthiness across claims, new business, and renewal underwriting.
Historically, claims follow one of two paths: the regular process for the vast majority, and the costly SIU process where hard fraud is suspected, or costs are well above norms. Both handle trustworthy and non-trustworthy claimants alike. The regular process is unnecessarily intense for honest claimants and not intense enough for the rest, because there has been no practical way to identify soft fraud.
With Clearspeed, two paths branch to four: high-risk indicators flag potential fraud where there was no SIU flag (~20% of claims), while low-risk indicators expedite claims that previously triggered rigorous follow-up. The result — capture up to 20% more fraud while moving honest claimants through faster.
| Clearspeed indicator | Distribution | Loss leakage / claim | CR benefit |
|---|---|---|---|
| Low risk of soft fraud (no SIU) | 75% | 0% | 0.8% |
| High risk of soft fraud (no SIU) | 20% | 30% | 3.4% |
| SIU-indicated claims | 5% | — | 0% |
| Total | 100% | 4.2% |
Assumptions: 20% of claims involve soft fraud; 30% loss leakage per soft-fraud claim; a low-risk indicator drives 10% expense savings; a high-risk indicator on a soft-fraud claim drives a 10% increase in investigation cost. Book: 60% loss, 10% LAE, 26% general expense, 4% profit.
The Clearspeed technology is game-changing in what it can do in the claims validation cycle — validating claims quicker than ever, with a confidence level you've never seen before.
02 — The cost · continued
The variables used to underwrite new auto business carry varying accuracy. Over 25 years, insurers pushed variables toward higher accuracy. This was largely through third-party data, but low-accuracy variables remain, driving premium leakage.
Clearspeed's voice questionnaire helps insurers validate information provided by applicants. It works inside any traditional or digital workflow: undisclosed drivers, rideshare or delivery use, annual mileage, violations, etc. Trustworthy applicants can be accelerated, reducing spend on costly third-party data: higher accuracy at lower cost.
| Clearspeed process | Distribution | Leakage avoided | CR benefit |
|---|---|---|---|
| Low risk of misrepresentation | 80% | 0% | 0% |
| High risk of misrepresentation | 15% | 20% | 3% |
| Application rejected | 5% | — | 0% |
| Total | 100% | 3% |
Assumptions: ~15% of accepted applications contain inaccurate or misrepresented information; insurers undercharge premium by ~20% on those. The opportunity is recapturing that undercharged premium — e.g. undisclosed drivers or personal vehicles used for ride-share/delivery.
02 — The cost · continued
Finding potential fraud at renewal underwriting is an unsolved problem in personal auto. Mostly insurers pull an MVR every three years or are reviewing poor-loss-ratio agents.
Why does this persist? Insurers aren't bothered because loss ratios historically improve as years-insured increase, and they worry about disturbing a more-profitable renewal book they can't segment. Mid-term endorsements create a natural touchpoint where a Clearspeed questionnaire can identify the policies that need re-underwriting.
| Clearspeed process | Distribution | Leakage avoided | CR benefit |
|---|---|---|---|
| No mid-term endorsement during year | 50% | 0% | 0% |
| Endorsement · low need to update | 37.5% | 0% | 0% |
| Endorsement · high need to update | 12.5% | 20% | 2.5% |
| Total | 100% | 2.5% / yr |
Assumptions: 50% of renewal policies undergo a mid-term endorsement; ~25% of renewal submissions carry data inaccuracies (persisting plus newly arising); insurers undercharge premium by ~20% on inaccurate renewals.
03 — The mechanism
Policyholders respond to a short series of yes/no questions via an automated questionnaire embedded in existing workflows. Clearspeed assesses the verbal responses for indicators of risk, and scores them from low-risk to high-risk in near real-time.
The proprietary risk model has been validated in both military and commercial environments, identifying risk based on vocal characteristics universal to all human beings. Insurers move trustworthy individuals forward with confidence, while high-risk cases are investigated further.
The evaluation is objective and unbiased. It uses no historical data and no personal identifiers, and works regardless of language, gender, race, location, credit score, or age. It triages and flags; it never decides. The organization receives the risk indicator; the individual never sees it.
I think that trust definitely gets eroded when customers do everything right, but they're still penalized. The customer did everything right, but the system failed. We need to stop asking customers to adapt to the limitations in our systems and start building systems that adapt to our customers.
04 — Resistance, adoption & adverse selection
The industry always adopts change of this size at scale. Leakage holes close in early-adopter carriers while staying open in laggards. Then, non-trustworthy customers shift from adopters to laggards, so the laggard's added cost equals the adopter's gain. The two effects compound.
Loss and premium-leakage holes close. Over two years the advantage grows to 8.8 points.
Laggards absorb the non-trustworthy customers shed by adopters, inheriting an equal and opposite cost.
The gap between adopters and non-adopters. At this size, it is game over for non-adopters. Eventually every carrier adopts, or fails.
05 — The prize
The total opportunity across the personal auto lifecycle. The same impact applies to other lines. Clearspeed is already working across home, life, and workers' compensation.
| Use case | Annual benefit | Combined ratio impact |
|---|---|---|
| Claims | $14.5B | 4.2% |
| New business underwriting | $2.1B | 0.6% |
| Renewal business underwriting | $13.8B | 4.0% (two years) |
| Total · $344B industry | $30.4B | 8.8% |
The recommendation
A carrier can survive inaction for a quarter. It cannot survive inaction forever.
Download PDFTrustFaster