In the United States financial services sector, no component is of course more personal than life insurance and underwriting.
As an Internal Medicine physician for 20 years, overlapping with 25 years as a life insurance medical director, I have seen both sides of the patient-consumer and underwriting business equation. What is the phrase? Up close and personal.
The business side is making significant gains in risk and mortality through generative AI and data retrieval from the EHR to impact speed, cost and accuracy, but system integration must keep pace along with navigating regulatory standards.
Two new papers look ahead
Two recent industry papers from March and August of this year assess the progress and posit the future. What’s missing though is whether the technological gains and emerging priorities of underwriting are also improving what the consumer - the insured policyholder - experiences.
RGA’s Underwriting 2050 paper, drawing on survey responses at the 2026 AHOU conference, found broad agreement that data will become more central, AI and automation will expand, and underwriters will focus more heavily on complex cases, exceptions, governance, and judgment, where “human expertise” must remain. From my experience on the carrier side - where I reviewed more than 5,000 cases per year - current automated approaches still often underestimate the complexity of medical assessments. Faster data retrieval does not automatically translate into better decisions. Interpretation is the constraint that needs further unlocking.
It’s equally important that all stakeholders advance best practices together. A recent Equisoft survey found that 38 percent of carriers “cite legacy IT systems lacking automation capabilities as their primary challenge in expediting underwriting decisions.” In other words, they are not yet ready for full artificial intelligence systems. The paper ends by stating the underwriter’s role will become more specialized toward greater trust and accountability. Here again I would bring this back to the consumer. Can they trust a policy and premium derived in a matter of hours?
Munich Re’s Leading with EHRs study shows what better evidence sequencing can accomplish. Having done work starting 10 years ago to establish EHR retrieval for life insurance capabilities, then exploring structured data to feed automation, I support this approach and refinements that have followed.
The study found that leading with the EHR increased the decision rate from 66 percent to 86 percent and improved modeled mortality protection compared with a traditional accelerated-underwriting program, indicating potential savings across evidence, underwriting time, and modeled mortality. Yet only 52 percent of the study cases could be decided without ordering additional medical billing data, and returned records vary in whether they are underwriting-ready. In other words, retrieving an electronic record is not the same as understanding it.
As a physician, that does not surprise me. A sharable record may document diagnoses, medications, and encounters without fully showing when a condition began, whether it is controlled, how it has changed, how it responded to treatment, or how multiple conditions interact. More data does not automatically create a coherent clinical picture. Here again interpretation is key.
Leading with the consumer
For consumers the process can still feel distant. They authorize access to records, may complete an examination, answer follow-up questions, and wait — often without knowing what information is missing or why more evidence is being requested. Acknowledging the middle market, more non-traditional families and dual-income households and that today’s patients have more than one primary care provider as a source of records, can result in unmet insurance needs by those who would benefit most from financial protection products. Technology can reduce that friction, but only if progress is measured not just in carrier efficiency, but in what the applicant actually experiences.
With often just a few consumer-to-insurer touchpoints per year, trust and loyalty, along with technological gains bringing accuracy, is another stakeholder goal among carriers, reinsurers and brokers. As the ability for consumers to query and access their own health records from data exchange networks increases, who will they choose to share it with?
Better use of available evidence should reduce unnecessary requirements, repeated requests, and avoidable delays. It should make the process clearer. Consumers should also have a reasonable opportunity to understand what information they are authorizing and how it will be used as underwriting catches up to other financial services sectors such as banking.
The balanced path forward
Artificial intelligence will be part of this future, but AI is not the final decision-maker. Technology can help organize evidence, surface unresolved questions, and support more consistent review. Underwriters, medical directors, carriers, and reinsurers must remain accountable for how that information is interpreted and applied.
These questions are part of why I have joined Scott Coller, PharmD, CEO and founder of Gesund Health Data. In the mid-2010’s, Scott championed interpretation as a key to underwriting rules engines. And he teaches Applied AI in his home state of Wisconsin. We share a belief that technology should make the clinical picture clearer while leaving underwriting judgment where it belongs — with underwriters, carriers, and reinsurers.
This is the beginning of the conversation, not the conclusion. As underwriting becomes more data-driven, the industry should continue asking not only how technology can make the process faster, but how it can make the process more understandable, trustworthy, and beneficial for the person behind the data.
Steven Dunlap, MD, is chief medical and innovation officer at Gesund Health Data.
Sources:
- RGA, Underwriting 2050: From Super Session to Strategic Foresight (August 2026): RGA’s PDF. rgare.com
- Equisoft and the Life Insurers Council survey: Equisoft’s report summary. It reports that 38% of respondents cited legacy IT systems lacking automation as a primary challenge to expediting underwriting decisions. The survey focused on small and midsize carriers and covered one-third of LIC members, so the piece should retain that scope rather than present it as an industrywide figure. equisoft.com
- Munich Re, “Leading with EHRs” (March 31, 2026): Munich Re’s study. It reports findings from a 144-application study subset with valid EHR data. munichre.com