What Is Straight-Through Processing in Underwriting? How Close Are We?

Straight through processing in underwriting sounds like the endgame: an application comes in, data gets pulled, a model scores the risk, and a policy goes out the door. No human touches the file. No one reviews the lab results or calls the attending physician. The whole thing happens in minutes.

Some carriers are getting close. Most are not.

The gap between the STP vision and operational reality tells you a lot about where life insurance underwriting actually sits in 2026. It also reveals which problems are genuinely hard and which ones the industry has been slow to solve for reasons that have nothing to do with technology.

According to Datos Insights (formerly Aite-Novarica Group), more than 80% of personal lines and individual life insurers have implemented at least some level of automated underwriting. But "some level" is doing heavy lifting in that sentence. Full STP--application to policy with zero human intervention--remains the exception, not the rule.

What straight through processing actually means in underwriting

STP originated in financial services, where securities trades needed to settle without manual reconciliation between counterparties. The insurance industry borrowed the term, but the meaning shifted. In underwriting, STP refers to the ability to receive an application, assess risk, make a decision, and issue a policy entirely through automated systems.

That sounds simple. In practice, it requires every link in a long chain to work without breaking: data ingestion from multiple third-party sources, rules engine evaluation, predictive model scoring, automated decision-making within approved authority limits, and electronic policy generation. If any single step needs a human, it's not STP.

Swiss Re's research on AI-driven predictive underwriting frames the core tradeoff clearly. Their analysis compares STP rates against model error rates--push automation too aggressively and misclassification increases. Pull back for accuracy and throughput drops. Finding the right balance depends on a carrier's risk appetite, product mix, and regulatory environment.

How the STP pipeline works

The typical automated underwriting pipeline in 2026 follows this sequence:

1. The applicant submits a digital application, usually through a direct-to-consumer portal or agent-assisted platform.
2. The system pulls data from prescription databases (Milliman IntelliScript is the most common), motor vehicle records, MIB, credit-based scores, and increasingly electronic health records.
3. A rules engine evaluates the data against the carrier's underwriting guidelines. Age, face amount, health indicators, and behavioral signals all factor in.
4. Predictive models assign a risk score. These models are trained on historical mortality data and incorporate variables like prescription patterns, chronic condition indicators, and demographic factors.
5. Applications that score within predefined thresholds receive automated approval. Those outside the threshold get routed to a human underwriter.
6. Approved applications move to electronic policy generation and delivery.

The whole thing can take under ten minutes for applications that stay on the automated path. Damco Solutions reported that AI-driven underwriting systems have cut decision times from five days to 12.4 minutes for standard policies, though that figure represents best-case scenarios at carriers with mature implementations.

STP rates across insurance lines

Not all insurance products lend themselves equally to automation. The complexity of risk assessment varies enormously by line of business, and STP rates reflect that.

Line of business	Approximate STP rate (2025-2026)	Why it's there
Personal auto	70-85%	Standardized risk factors, deep actuarial data, commoditized product
Term life (low face amount)	50-65%	Accelerated underwriting programs handle simpler cases well
Term life (high face amount)	15-25%	Larger policies trigger more conservative review thresholds
Whole life / permanent	10-20%	Product complexity and higher face amounts limit automation
Small commercial	40-55%	Standardization efforts paying off, but variability remains
Large commercial / specialty	Under 10%	Too many unique risk factors for rules-based automation
Group life	30-45%	Employer-level underwriting is simpler than individual

These numbers come with caveats. Carriers define "STP" differently. Some count any application that doesn't require a full manual review. Others only count cases with zero human intervention from submission to policy delivery. The Datos Insights 2023 report (their most recent comprehensive survey) found that definitions varied enough across respondents to make precise benchmarking difficult.

What's actually blocking full STP

The barriers are a mix of technical limitations, regulatory constraints, and institutional inertia. Some are getting solved. Others are structural.

Data quality and availability. STP depends on clean, complete, real-time data feeds. Electronic health records are still fragmented across health systems, and many applicants don't have the digital health footprint that automated models need to score confidently. When the data is thin, the system has to escalate to a human.

Fraud detection gaps. CRL Corp has estimated that tobacco misrepresentation alone costs the life insurance industry roughly $4 billion annually. Without fluid verification (blood and urine tests), carriers lose a direct fraud detection mechanism. Any STP system needs alternative ways to catch misrepresentation, whether through prescription history analysis, lifestyle data, or biometric signals.

Model confidence thresholds. Swiss Re's work on predictive underwriting highlights a fundamental tension: higher STP rates come with higher error rates. A carrier that automates 80% of decisions will misclassify more applicants than one that automates 40%. The acceptable error rate depends on product pricing margins, reinsurance terms, and regulatory scrutiny.

Regulatory requirements. Some jurisdictions still require specific underwriting documentation or human sign-off for certain policy types. State insurance departments in the U.S. have varying positions on algorithmic underwriting, and several have issued guidance requiring explainability for automated decisions.

Reinsurer comfort level. Reinsurers who price treaties need confidence that automated underwriting produces mortality experience within expected bands. Munich Re's Biometric Portfolio Analysis platform, built on data from more than 30 participating insurers over 15 years, provides some of that analytical infrastructure. But reinsurers remain cautious about cohorts underwritten without any biometric verification.

Where contactless biometrics fit in

One of the persistent problems with pushing STP rates higher is the loss of biometric data. When you eliminate the paramedical exam, you lose objective health measurements--heart rate, blood pressure, oxygen saturation--that help validate what applicants report on their applications.

This is where camera-based health screening enters the picture. Remote photoplethysmography (rPPG) technology can capture vital signs through a smartphone camera during the application process itself. The applicant does a 30-second scan, and the system gets real-time biometric data without scheduling a nurse visit or mailing a test kit.

For STP specifically, contactless vitals address two problems at once. They provide the biometric data that predictive models need for higher-confidence scoring, and they add a layer of fraud detection (it's harder to misrepresent your health when the system is reading your actual vital signs). Circadify has developed this capability for the insurance vertical, designed to plug into existing digital application workflows. More details on how this fits into the underwriting stack are available at circadify.com/industries/payers-insurance.

How carriers are approaching STP implementation

The carriers making real progress on STP share a few common traits. They aren't trying to automate everything at once. Instead, they're segmenting their book and automating the portions where data availability and model confidence are highest, then gradually expanding the automated path.

The segmentation approach

Gen Re's 2024 Individual Life Accelerated Underwriting Survey found that 57% of total individual life applications were eligible for accelerated processing. Of those, about 20% received fully automated decisions. The remaining 37% went through an accelerated path but still involved human review at some point.

That 20% figure is roughly where the industry's STP frontier sits for individual life. The next increment--getting from 20% to 40% or 50%--requires solving the data and fraud problems discussed above.

What the technology stack looks like

Carriers with high STP rates typically run three layers of decisioning:

A rules engine handles the first pass, filtering applications based on hard eligibility criteria (age limits, face amount caps, state availability). A predictive model layer scores risk for applications that pass the rules filter. And a decision authority layer determines whether the model's output falls within the carrier's automated approval threshold or needs escalation.

The third layer is where most STP initiatives stall. Setting the automated approval threshold involves actuarial sign-off, reinsurer agreement, and regulatory review. It's a governance problem as much as a technology problem.

Current research and evidence

The empirical base for automated underwriting continues to grow, though long-term mortality studies on STP cohorts are still limited.

RGA and the University of Leicester published research using UK Biobank data from 407,569 participants showing that wearable and biometric health data--resting heart rate, physical activity, sleep patterns--can "dramatically improve the ability to differentiate mortality and morbidity risks." This matters for STP because it suggests that non-traditional data sources can compensate for the loss of traditional lab panels.

Gen Re's 2024 survey remains the most comprehensive industry benchmark, documenting that 82% of U.S. life carriers now have fully or partially implemented accelerated underwriting workflows, up from 62% planning implementation in 2019.

Datos Insights has tracked STP rates across insurance lines since 2019, showing steady but incremental progress. Their finding that more than half of personal lines, small commercial, and individual life insurers process "substantial volumes" through STP indicates the baseline is established. The question now is expansion.

The future of straight through processing in underwriting

Full STP for the majority of life insurance applications is probably still three to five years away. The technology exists for simple cases, but the convergence of better biometric data, more mature predictive models, and regulatory clarity all need to come together.

A few developments are worth watching. Real-time biometric capture during the application process could push automated approval rates significantly higher by giving models more input data. Large language models are beginning to handle unstructured medical records (attending physician statements) that historically required human review. And the younger applicant demographic, with more digital health data available, is naturally easier to underwrite algorithmically.

The carriers that reach 50%+ true STP first will likely be the ones that solve the biometric data problem--getting objective health measurements into the digital application workflow without adding friction. That's the bottleneck, and it's the one where the most interesting work is happening right now.

Frequently asked questions

What does STP mean in insurance underwriting?

Straight through processing (STP) in insurance underwriting means an application is received, risk-assessed, decided, and issued as a policy without any human intervention. The entire process runs through automated systems, from data collection through policy delivery.

What percentage of life insurance applications are processed via STP?

It depends on the product type and carrier. For low face amount term life, roughly 50-65% of eligible applications go through automated paths, though only about 20% achieve true zero-touch STP. Personal auto insurance has the highest rates at 70-85%. Complex products like whole life and large commercial policies remain largely manual.

Why can't insurers automate all underwriting decisions?

Several factors limit full automation: incomplete digital health data for many applicants, fraud detection challenges without biometric verification, regulatory requirements for human oversight in some jurisdictions, and reinsurer concerns about mortality experience in fully automated cohorts. Model confidence thresholds also mean that unusual or borderline cases still need human judgment.

How does contactless health screening help increase STP rates?

Contactless screening through rPPG technology captures real-time vital signs (heart rate, blood pressure, oxygen levels) via smartphone camera during the application. This gives automated models more biometric data to work with, improving scoring confidence and reducing the need for human escalation. It also provides a fraud detection layer by verifying actual health status against self-reported information.