The carriers who gain an edge won’t be the ones with more data, but the ones with a decision-first framework for using it.
Consider the speed at which various underwriting assets arrived in the 20th century: Underwriters opened each new case knowing the applicant’s age and sex and had access to some health information in the form of self-reported application data. But when more information was needed—as it usually was—marking progress on most cases required a calendar.
Attending physician statements were detailed, but they were expensive, slow to arrive, and often difficult to digest. Paramedical exams and insurance labs were somewhat faster, but still invasive, costly, and limited in scope. Underwriters learned to assemble the case piece by piece because that was the only practical way to manage uncertainty.
The entire process prioritized accumulating evidence and building a rationale for defensible decisions. It was cumbersome, but customers, brokers, and agents all expected the process to take weeks, so no one was at a competitive disadvantage. Many On the Risk readers work for companies that have been insuring Americans for over a century, so that approach obviously worked for a long time.
A new kind of adverse selection
Today, underwriting departments are under pressure to meet customer expectations for quick, seamless transactions. But while applicants increasingly expect life insurance transactions to feel like other digital customer experiences, the consequences of making a bad decision aren’t going away for underwriters.
Meanwhile, new sources of health and consumer data are available at a speed and scale the industry couldn’t have imagined even ten years ago.
These assets change the underwriter’s relationship to evidence: the question is no longer always, “What else do I need to defend a decision?” but rather, “Have I already seen enough to decide?”
The fact that new customer expectations and new data sources and tools are hitting the market at the same time is significant. Now, companies that don’t make good use of them are going to be penalized in the market.
This creates a new kind of adverse selection. That term previously only applied when a carrier insured someone that competitors who knew better would decline. Now, decision speed has become part of risk selection. If two carriers would ultimately reach the same underwriting conclusion on an insurable applicant, the carrier that gets there a few days late is at a disadvantage.
Just digitizing legacy processes isn’t the answer
Underwriting with a legacy mindset meant assembling the case: ordering more evidence until the rationale felt defensible. That “all the data, all the time” approach made sense when information was scarce and guidelines existed largely to tell underwriters what else to order.
Digitizing legacy processes may slightly streamline things, but let’s face it: it was not the United States Postal Service’s fault that it took a month or more to get an APS; now they arrive as .pdf attachments and they still take weeks.
Nor does the mere fact that assets arrive digitally make a process “modern.” Proposals to start every case with an EHR overlook comparatively low hit rates, hours or days-long waits, and significant underwriter touch time. They can be so long and repetitive that a cottage industry has sprung up to offer summaries prepared by generative AI.
A truly modern approach starts with the decision: identify the risk pattern in the evidence already available and seek more documentation only when it is likely to change the outcome.
Defining the decision-first mindset
The life insurance industry spends a lot of time debating automation versus traditional underwriting. The better question is not whether a manual search might uncover more evidence, but whether additional evidence is worth its cost once the risk pattern is already clear.
Life underwriters are not paid to collect facts. They are paid to make sound decisions about uncertain futures. A decision-first mindset means knowing when the available evidence is sufficient to reveal the underlying risk pattern. The goal is not to eliminate every uncertainty, but to identify the uncertainties that matter. The fastest route to a decision begins with acquiring the relevant underwriting assets that are instantly available, indexed, and interpreted.
“Instant” is self-explanatory; “indexed” refers to data that arrives structured, making it suitable for easy integration into automated systems; “interpreted” is meant in the sense that facts arrive with a risk score and/or clinical conclusions already drawn. Such assets include prescription histories, medical claims data, lab test histories, MIB data, and credit attributes.
This approach can collapse time-to-decision in many cases. Sophisticated rules engines and predictive models give underwriters more than a list of impairments; they offer insight into severity, condition management, protocol adherence, and comorbidities. Often, that is enough to support a sound decision at or near the point of application.
Some cases will still require an EHR, APS, or other follow-up. But modern evidence should help determine when those records are truly needed, which records matter, and what the underwriter should be looking for when they arrive.
It’s time to rethink guidelines
Many underwriting guidelines were shaped when an impairment might be suspected early, but its severity or management remained unclear. When comprehensive evidence is already available, structured, and interpretable, guidelines should help answer a different question: What, if anything, would materially improve the decision?
Instead, we’re still ordering EHRs or APSs whenever we see certain impairments. Especially in accelerated underwriting products, we may jet-decline or rate such applicants, exposing us to opportunity costs. For example, an underwriter might see the anticonvulsant divalproex (Depakote) in an applicant’s prescription history—that being one instant data type that is already nearly ubiquitous. Divalproex may be prescribed for migraines, epilepsy, or bipolar disorder. Seeing this could trigger a knee-jerk EHR or APS order to determine whether the underlying diagnosis is insurable, but it’s entirely likely that determination could also be made faster and cheaper, if the underwriter just had access to the applicant’s medical claims data from the outset.
Underwriters, many of whom take every early-death claim personally, have historically found comfort in guidelines, too; sticking to them provided protection for their careers. But those old habits prevent us from fully exploiting the complete and nuanced data that’s now available from the moment of application.
A good first step is to begin with the most comprehensive evidence available at the point of application and, when it supports a sound decision, end the process there. That enables a 21st-century customer experience and avoids anti-selection on the basis of decision time.
Guidelines still matter, but guidelines built for different and often incomplete information should no longer automatically dictate workflows. In many cases, the instantly available evidence is sufficient; in some, the cost in time or money associated with seeking additional records is justified when the ensuing decision is changed. We need to know the difference.
The next evolution in underwriting will not come from digitizing every old step and collecting more information. It will come from rethinking which steps materially improve the decision.
Chris Martin, Principal and Managing Director, received a B.A. in English from the University of Illinois and an MBA from the University of Chicago Booth School of Business.
This article was originally featured in the September 2026 Issue of OTR and is reprinted with permission of ON THE RISK, Journal of the Academy of Life Underwriting (www.ontherisk.com).
