Catalog record · Operations & Efficiency
Stage as of Nov 5, 2026 · historical report, not live reconfirmation.
AI-SDLC agents across insurance software delivery
Vendor workflow source. A reviewed supplier account names the customer and workflow. Supplier attribution and deployment claims are not independent outcome validation.
Last researched Nov 8, 2026. Source access and reported stage are recorded below; deployment evidence does not establish ROI.
What the research supports
Verisk uses five agents for requirements, grooming, development, code review and QA inside its engineering tools, with human approval between stages.
- Reported stage
- Production
- Technology
- GenAI/LLM
- Business function
- Back Office
- Impact (legacy label)
- Not labeled
- Catalog date · unconfirmed
- Not reported
- Imported
- Nov 8, 2026
- Last researched
- Nov 8, 2026
Measures & evidence
20–30 · percent
Overall cost per story point reduction
Measurement dates unstated; reported 2026-11-05
Scope, baseline & source
- Scope / denominator
- Features developed end-to-end with all five AI-SDLC agents
- Baseline
- Not disclosed in this account
- Claim basis
- AWS/Verisk coauthored case
Includes labor plus AWS compute; selection method and sample size undisclosed.
Read claim source ↗>90 · percent
Output acceptance
Same reported implementation
Scope, baseline & source
- Scope / denominator
- Outputs across the complete workflow
- Baseline
- Not disclosed in this account
- Claim basis
- AWS/Verisk coauthored case
Human reviewers accepted with minimal modification; not defect-free code.
Read claim source ↗3–4 hours → under 10 minutes · time per test-planning activity
Test-plan preparation time
Same reported implementation
Scope, baseline & source
- Scope / denominator
- Features developed end-to-end
- Baseline
- Not disclosed in this account
- Claim basis
- AWS/Verisk coauthored case
Task time, not net financial savings.
Read claim source ↗Measures retain their original currency, scope and period. Capacity, adoption and forecasts are not realized cash savings.
Documented vendor relationships
No vendor relationship has been linked to this specific workflow in the curated overlay. This does not mean it was built without vendors.
See the company’s documented relationshipsDeployment evidence & limits
Primary source describes actual use; publication date is not necessarily the original go-live date. Status as of 2026-11-05.
Stage is as reported in the cited evidence. Current operating status has not been independently reconfirmed; the research date is an editorial review date.
One-week releases are a target, not demonstrated frequency. Exact original deployment date and independent evaluation absent.
Recorded dependencies
- Human review, ticket/repository context and auditable tool permissions.
Source register
- Reinventing software delivery with AI-SDLC on Amazon Bedrock AgentCoreaws.amazon.com · opens in a new tab
AWS with Verisk engineering coauthors · vendor-case · full-text · Published 2026-11-05
Research lanes: frontier-refresh · Batch research-refresh-2026-11-08. Coverage and remaining gaps