Data in
Broker email intake
Applications, schedules, loss runs
← Solutions Lab
Case study · Insurance
Submission triage for a mid-sized commercial insurer
Client
A mid-sized US commercial property and casualty insurer
Industry
BFSI
Solution families
AI and Autonomous Operations · Data and Decision Intelligence
Time to proof
4 weeks
Executive summary
Underwriters received far more broker submissions than they could review, so good risks waited in the inbox. QAO SubmissionIQ reads every submission, checks it against the insurer's appetite and ranks the queue, so each day starts with the best opportunities and a one-page summary for each.
Key outcomes
70%
Reduction in triage time
3x
Submissions reviewed per underwriter per day
100%
Submissions screened against appetite on arrival
Solution and architecture
What we composed.
Broker emails flow into SubmissionIQ, which extracts the insured, location, property values and loss history, checks them against the appetite rules and scores each submission. Underwriters work a ranked queue with a one-page summary where every fact links to its source, and decisions flow back to policy administration with a full audit trail.
Intelligence
Extraction and appetite agent
Reads, checks, scores
Workflow
Underwriter workbench
Ranked queue and summaries
Systems of record
Policy admin and CRM
Decisions and audit trail
Solutions Intelligence, applied.
Level 1
Sense
Mapped the submission flow and measured how long each review took.
Level 2
Reason
Turned the appetite in underwriters' heads into written rules.
Level 3
Compose
Designed extraction, scoring and the workbench around the existing inbox.
Level 4
Prove
Ran on a month of synthetic submissions against agreed accuracy targets.
Level 5
Evolve
Added lines of business and tuned scoring from underwriter feedback.
Technology
PythonDocument AILLM gatewayRules enginePostgresReact