Insurance / Intelligent UnderwritingMercury Insurance2026-08-08

Mercury Insurance: Clean Submissions Decided in Minutes

Underwriting personal auto or homeowners means assembling evidence from many places — the application, motor vehicle and loss-history reports, prior carrier declarations. Mercury built intelligent underwriting on MOI: submissions are parsed into structured facts, scored against state and product guidelines, and clean ones decided automatically.

Mercury Insurance

Mercury General Corporation, founded in 1961 and headquartered in Los Angeles, is a US multiple-line personal insurer writing primarily personal auto plus homeowners through 12 subsidiaries across 11 states. California accounts for roughly 82% of direct premiums written, where Mercury is the fourth largest private-passenger auto insurer.

8,500+
Independent agents
Minutes
Clean submission decisions
State + product
Guideline scoring
Multi-source
Reports and images parsed

The challenge

Underwriting a personal auto or homeowners policy means assembling evidence from many places: the application, motor vehicle and loss-history reports, prior carrier documents, vehicle and property data, inspection photographs, and third-party and geospatial risk data such as wildfire exposure. Most of it arrives as documents and images rather than clean fields. Underwriters re-key values, chase missing evidence and apply guidelines that differ by state and product; outcomes vary between reviewers, referral queues build at peak-and a submission that stalls with an agent is a submission that goes to another carrier.

The solution

Mercury built intelligent underwriting on OmniFabric. Submissions and their attachments are ingested and parsed-applications, reports, prior declarations pages and inspection photographs-with OCR, table extraction and vision models turning documents and images into structured facts. Those facts are validated against internal and third-party data, then scored against the applicable state and product guidelines, with risk features computed from the same governed data pricing and actuarial teams use. Clear submissions are decided automatically within appetite; the rest reach an underwriter as a referral carrying the exposure summary, the rule that triggered and links to the underlying evidence.

The outcome

With around 8,500 independent agents producing the great majority of premium, turnaround is competitive advantage: clear submissions are decided without waiting in a referral queue.

The same rules run against every submission in every state, so an outcome no longer depends on which underwriter picked it up or how busy the queue was that day.

Routine submissions clear automatically, concentrating underwriter judgment on genuinely complex or borderline exposures instead of on data entry.

Solution Architecture

Data sources
  • Applications & Agent Submissions
  • MVR, Loss History & Prior Carrier Documents
  • Vehicle, Property & Inspection Photographs
  • Third-Party and Geospatial Risk Data
MatrixOne Intelligence
  • Document and Image Parsing into Structured Facts
  • Validation Against Internal and Third-Party Data
  • Guideline Scoring by State and Product
  • Risk Features from the Same Governed Data as Pricing
  • Decision, Evidence and Outcome Retained for Audit
Business applications
  • Straight-Through Decisions Within Appetite
  • Referrals with Exposure Summary & Evidence
  • Structured Risk Record per Submission
  • Audit Trail Behind Every Decision

Related solutions and products