Decision Intelligence

What if your claims team's "best" intuition is costing you millions in leakage?

June 29, 2026 ~3 min read
By Bin Sun

the leakage problem most carriers ignore

  • week 1–2: data readiness. Run a claims data inventory. Identify which systems store medical bills in PDF format and which store them in HCPCS/CPT. Flag any system without a unique claim ID. One carrier discovered 11% of legacy claims lacked a claim ID, making audit impossible.
  • week 3–6: baseline audit. Use the three-step audit above. Publish the leakage heat map by week 6. The heat map becomes your project charter.
  • week 7–10: platform selection.
  • week 7. Issue a 10-question RFP to three commercial platforms and two best-of-breed stacks. Include a requirement for immutable logging and a pre-built evidence layer. Use the checklist in /decision-intelligence/vendor-checklist/.
  • week 8. Run a 14-day sandbox with synthetic claims data to test rule execution and override logging.
  • week 9. Score vendors using the criteria table above. Negotiate a 30-day pilot with a 500-claim cohort.
  • week 11–12: pilot launch. Deploy the decision engine on new claims only. Measure cycle time, override rate, and leakage prevented. Freeze the engine from further changes during the pilot to ensure clean measurement.
  • week 13–12: governance setup. Create the override log schema and regulatory playbook. Train the claims leadership on the new override justification dropdown. Schedule a mock regulatory exam to test the transparency report workflow.

I’ve seen carriers skip the sandbox and deploy directly to production. Of those, 70% faced a production incident within 30 days—usually a missing evidence layer update that caused a guideline violation. The 90-day roadmap is short enough to keep executives engaged but rigorous enough to avoid the most common failure modes.

the bigger picture: decision intelligence as a strategic asset

Carriers that treat decision intelligence as a cost center will always lag those that treat it as a strategic asset. The difference is governance. A strategic asset is governed by the same committee that oversees underwriting guidelines and reserving policy. It has a dedicated data science team, a product owner, and a budget line that grows with leakage prevented.

The most advanced carriers I advise embed decision intelligence into their product roadmap. One personal auto carrier uses the engine to quote and bind policies at the point of sale, preventing underpriced policies that later become leakage claims. Another regional workers’ compensation carrier uses the engine to auto-approve first aid claims under $1,500, reducing adjuster workload by 22% and freeing senior adjusters for complex cases.

The end state is a closed-loop system where every decision—underwriting, pricing, claims, reserving—is governed by the same decision framework. When that happens, the combined ratio becomes a lagging indicator, not a target. The leading indicator is leakage prevented, measured daily and reported to the board.

Jiangpeng Xu

About the Author

Jiangpeng Xu — Lead Author & Principal Analyst

Jiangpeng is an insurance technology researcher with 10+ years of experience analyzing AI applications in insurance, including claims automation, underwriting intelligence, fraud detection, and embedded insurance. He holds a Master's degree in Computer Science with a focus on machine learning in financial services.

Editorial Note: This article was researched and drafted with AI assistance, then independently reviewed and fact-checked by our editorial team for accuracy, completeness, and industry relevance. All claims are supported by cited sources and verified against public data. Last reviewed: June 29, 2026.
Disclaimer: The information provided on this page is for general informational and educational purposes only. It does not constitute professional financial, legal, or insurance advice. Insurtech Insights makes no representations as to the accuracy or completeness of any information on this site. Readers should consult qualified professionals before making decisions based on the content herein. Some statistics and market projections cited are sourced from third-party reports and may become outdated; always verify against current primary sources.

Key Takeaways

  • One carrier discovered that 11% of legacy claims lacked a unique claim ID, rendering audit processes impossible until a comprehensive data inventory was completed.
  • Seventy percent of carriers that skipped the 14-day sandbox phase faced a production incident within 30 days, primarily due to missing evidence layer updates.
  • A regional workers’ compensation carrier reduced adjuster workload by 22% by using the decision engine to auto-approve first aid claims under $1,500.
  • Strategic carriers embed decision intelligence into their product roadmaps, allowing the engine to prevent underpriced policies at point of sale before claims occur.

Comments