What problem does it solve?
Underwriting a life insurance application has traditionally meant forms, medical questions, exams, lab tests and a human underwriter reading the file, which takes days or weeks. Applicants drop out while they wait, and underwriters spend time on healthy applicants whose cases would have been approved anyway.
Insurers now use AI models, electronic health records and other data to decide the clean cases in seconds and keep human underwriters for the rest. The limits are set by the insurer and its reinsurer: the amount of cover, the age and the completeness of the data. Manulife, for example, says in its 2025 annual report that it raised the amount of cover that qualifies for an instant decision in the US from US$3 million to US$5 million, working with Munich Re Life US and its AI driven risk assessment platform alitheia.
How does it work?
- Collect the application. The applicant answers health and lifestyle questions online or through an agent, and consents to data checks.
- Enrich with data. The system retrieves prescription history, motor vehicle and other records, and electronic health records where the market allows.
- Score and decide. A risk model rates the application. Cases that fit the insurer's rules for cover amount, age and data quality are approved or priced automatically, and the decision comes back in seconds.
- Route the rest. Cases outside those limits go to an underwriter with the data and the model's reasons.
- Audit. The insurer samples automated decisions, compares mortality and claims with the model's expectations and adjusts the limits.
- Audience
- Back office
- Autonomy
- Supervised agent
- Adoption
- Early adopters
- Channels
- API and system to system
What is it worth?
Benchmarks are computed from the public deployments below: one data point per organization per KPI, with who made each claim.
| KPI | Median | Reported range | Data points | Claimed by |
|---|---|---|---|---|
| Automation rate | Too few to pool | 94% | 1 | 1 organization |
| Cycle time reduction | Too few to pool | about 50% | 1 | 1 organization |
Value drivers: Speed and cycle time, Lower cost to serve, Customer experience.
Indicative value
A life insurer with 40,000 new individual life applications a year
USD 400,000 to USD 2.4 million
Underwriting cost avoided per year
How this is calculated
Formula: applications * automatedShare * costPerCase. The low scenario uses every low input, the high scenario every high input.
| Input | Low | High | Basis |
|---|---|---|---|
| New individual life applications applications, applications per year | 40,000 | 40,000 | The reference insurer. |
| Share of applications decided straight through automatedShare, fraction of applications | 0.2 | 0.4 | Editorial assumption, kept well below the figures on this page: Ping An reports 94% of life policies underwritten within seconds, which measures speed rather than the share with no human review, and Sun Life reports a 50% increase in straight through underwriting without the base rate. Replace with the share your reinsurer approved limits allow. |
| Underwriting cost avoided per straight through case costPerCase, USD per application | 50 | 150 | Editorial assumption for underwriter time and medical evidence that is no longer ordered. Replace with your own fully loaded cost per underwritten case. |
What it leaves out: Gross underwriting cost only. It leaves out the data provider and model fees, the validation work, the revenue from applicants who no longer drop out while they wait, and any extra claims cost from anti selection in the automated group.
Who already uses it?
3 public deployments, strongest evidence first. Grades: A regulator or audit, B the organization itself, C vendor case study, D anonymous or estimate.
Ping An Insurance
China · Insurance · 2025
Ping An says on its technology page that it applies big data, AI and machine learning to digital underwriting of life insurance. For 2025 it reports that 94% of life insurance policies were underwritten within seconds. The page does not say what share of those policies were decided with no human review, which rules were applied, or how the figure is counted, so the number shows speed of a decision rather than a confirmed rate of fully automated decisions.
- Automation rate: 94%, Life insurance policies underwritten within seconds, 2025
"In 2025, 94% of life insurance policies were underwritten within seconds."
Claimed by: organization
Sun Life
Canada · Insurance · 2025
Sun Life's 2025 annual report says AI tools cut median response times for individual insurance applications in a target segment by close to half, including a 50% increase in straight through underwriting. In its Client Contact Centre, generative AI raised chatbot containment by 16 percentage points year over year, and advisors use a generative AI Notes Assistant Tool; in Hong Kong it launched Advisory Buddy, a GenAI chatbot inside its Advisor Workbench. In Malaysia almost two thirds of clients received automated underwriting decisions within two hours.
- Cycle time reduction: about 50%, median response time for individual insurance applications, target segment, 2025
"Leveraged AI tools to do more for our Clients including reduced median response times for individual insurance applications for a target segment by close to half (including increasing straight-through underwriting by 50%), increased chatbot containment rate (up 16 percentage points year-over-year) with generative AI tools in the Client Contact Centre, and enhanced advisor productivity using a generative AI-powered Notes Assistant Tool."
Claimed by: organization
Manulife
Canada · Insurance · 2024
Manulife's 2024 annual report says generative AI in Singapore automates document digitization and summarization in underwriting, improving the accuracy of underwriting decisions and reducing processing time for policy applications, and that in the US it expanded the use of electronic health records and used generative AI to automate preliminary underwriting assessments. In 2025 it partnered with Munich Re Life US on alitheia, an AI driven risk assessment platform, raising the instant underwriting decision eligibility limit from US$3 million to US$5 million. No time or accuracy figures are disclosed.
No outcome disclosed.
How do you implement it?
A model agnostic playbook: what to prepare, the order to build in, and what goes wrong.
Data you need
- Historical applications with underwriting decisions and later claims and lapse outcomes
- Reinsurer approved rules for the amounts and ages eligible for automatic decisions
- Consent and legal basis for each external data source
Systems to integrate
- Policy administration and new business system
- Electronic health record, prescription and other third party data providers
- Reinsurer platform or underwriting rules service
- Underwriter workbench for referred cases
Complexity: High
It needs the policy administration system, reinsurer approved underwriting rules, external health and records data sources with consent, a risk model validated against the insurer's own mortality experience, and an audit process for automated decisions.
- 1
Define the eligible population
Agree with the reinsurer which ages, cover amounts and products can be decided automatically, starting with the lowest risk segment.
- 2
Validate the model against your own experience
Run it in shadow mode against human underwriting decisions and later claims, and measure how often it would have approved cases an underwriter declined.
- 3
Launch for a narrow segment
Switch on automatic decisions for one segment, with every referral and a sample of automated decisions reviewed by an underwriter.
- 4
Widen the limits on mortality results
Raise cover amounts or ages only when claims and mortality in the automated group match the pricing assumptions.
Guardrails
- Automatic approvals only inside limits agreed with the reinsurer
- A path for the applicant to request human review of any decline or rating
- Sampling of automated decisions by senior underwriters every month
- Documented explanations for adverse decisions
- Consent recorded for each external data source
KPIs to instrument
- Share of applications decided automatically
- Median time from complete application to decision
- Mortality and claims in the automated group against pricing assumptions
- Referral rate and decline rate by segment
Human in the loop
Underwriters own every referred case and every decline, and review a sample of automated approvals. The chief underwriter and the reinsurer own the eligibility limits.
Common failure modes
- Anti selection
- Applicants who know they are ill learn which answers pass. Monitor mortality in the automated group and tighten limits when it exceeds pricing.
- Unfair outcomes
- A model trained on past decisions or on data that proxies for protected traits disadvantages some groups. Test results by group and keep adverse decisions explainable.
What are the risks and rules?
EU AI Act
High risk
Annex III point 5(c) lists AI used for risk assessment and pricing in relation to natural persons in life and health insurance as high risk, and Article 27 requires the insurer, as deployer, to carry out a fundamental rights impact assessment before first use. Because eligible applicants receive a decision with no human review, GDPR Article 22 on decisions based solely on automated processing also applies: the insurer needs one of the grounds in Article 22(2) and must offer human intervention and a way to contest the decision. The decision uses health data, so Article 22(4) in practice requires the applicant's explicit consent.
Rules that apply
Guidance
- Regulation (EU) 2024/1689 (AI Act), Article 6 and Annex III point 5(c) (European Union, Europe). Annex III point 5(c) names AI systems for risk assessment and pricing in life and health insurance as high risk.
- SB21-169, Protecting Consumers from Unfair Discrimination in Insurance Practices (Colorado Division of Insurance, North America). Requires insurers to test external consumer data, algorithms and predictive models for unfair discrimination; Regulation 10-1-1 sets governance requirements for life insurers that use them, as accelerated underwriting does.
Controls to put in place
- Model validation against the insurer's own experience before each limit increase
- Audit trail of the data, the score and the decision per application
- Monthly sampling review of automated decisions
- Fairness testing of approval and decline rates by group
- Fundamental rights impact assessment under Article 27 of the AI Act before first use
Frequently asked questions
- What share of life applications can be decided automatically?
Ping An reports that 94% of its life insurance policies were underwritten within seconds in 2025. Sun Life says AI tools cut median response times for individual insurance applications in a target segment by close to half in 2025, including a 50% increase in straight through underwriting. Both figures are the insurers' own, neither says how many cases had no human review at all, and both depend on the product mix and cover amounts.
- Is automated life underwriting high risk under the EU AI Act?
Yes. Annex III point 5(c) lists AI for risk assessment and pricing in life and health insurance for natural persons as high risk, so the insurer needs risk management, logging, data governance and human oversight.
- Which applications still go to an underwriter?
Large amounts of cover, older applicants, incomplete data and any case the model cannot rate with confidence. Insurers keep underwriters for these referred cases and for review of declines.
How to cite this page
Blits.ai AI Use Case Library, "AI for straight through underwriting decisions on life insurance applications", last verified 12 November 2026, https://www.blits.ai/ai-use-cases/life-straight-through-underwriting-decisioning. Licensed under CC BY 4.0. Method: how we verify use cases.
Changelog
- 12 November 2026: First published