Aviva to launch AI summarisation tool for life insurance underwriting
Aviva will roll out an AI-powered summarisation tool to its individual life insurance underwriting teams from 28 November 2025. The system uses generative AI to read and summarise GP medical reports so underwriters can assess cases faster with a clearer view of the facts.
After 18 months of testing and controls, the tool has handled 1,000 live cases and improved turnaround times during the active test phase. The aim is simple: reduce time spent on long digital medical reports while keeping decision quality tight.
Robert Morrison, chief underwriting officer at Aviva, said: "Protection underwriting in our industry has faced challenges for some time with underwriters needing to review lengthy digital medical reports, which can be extremely time consuming. The new summarisation tool will deliver real benefits for customers and underwriters alike. By leveraging generative AI responsibly, we're improving efficiency without compromising accuracy or care."
What this means for underwriting teams
- Faster case triage: Summaries surface key clinical history, dates, and outcomes so you get to a provisional view quicker.
- Sharper reviews: Reduced admin time on lengthy reports frees capacity for judgment calls, edge cases, and referrals.
- Consistency: Standardised summaries help align decisions across teams and reduce variance.
- Audit support: Structured outputs make it easier to document rationale and maintain oversight.
How the tool fits into the workflow
- Input: Digital GP reports are ingested once customer consent and data permissions are confirmed.
- Processing: The model extracts and summarises relevant medical details, timelines, and risk flags.
- Review: Underwriters use the summary as a starting point, with full documents available for verification.
- Decision: Human underwriters remain accountable for outcomes and exceptions.
Controls and governance (what to check)
- Data protection and consent pathways aligned to UK regulation (see the ICO's AI and data protection guidance: ICO AI guidance).
- Bias and error monitoring: Track agreement rates between AI summaries and final decisions; investigate variances.
- Human-in-the-loop: Clear thresholds for mandatory manual review and medical underwriting referral.
- Model risk management: Version control, change logs, and performance reporting to the risk committee.
- Customer impact: Complaint trends, TAT improvements, and fairness outcomes across demographics.
KPIs to watch post-rollout
- End-to-end turnaround time (initial review and total case cycle time).
- First-time right decisions and rework rate.
- Underwriter productivity (cases per FTE per week) without quality slippage.
- Referral rates to senior underwriters or medical officers.
- QA pass rates and adverse deviation on claims experience over time.
Impact for advisers and partners
- Potentially faster decisions and fewer follow-up queries for common medical disclosures.
- Clearer underwriting notes can help explain decisions and improve customer understanding.
- Case selection and pre-sales guidance may benefit from more predictable outcomes.
What this doesn't replace
Human judgment. Complex histories, conflicting evidence, or unusual risk patterns still need experienced underwriters. The tool speeds the review; it doesn't make the final call.
Practical next steps for insurance leaders
- Refresh underwriting guidelines to reference AI-generated summaries and escalation rules.
- Train teams on reading AI outputs, validating sources, and spotting hallucinations or gaps.
- Tighten documentation standards so every decision has a clear trail from summary to outcome.
- Coordinate with compliance and claims to align on fairness, explainability, and customer communications.
If you're planning similar adoption across underwriting or operations, upskilling your team on applied AI can shorten the learning curve. Explore role-based programs here: AI courses by job.
Your membership also unlocks:
