AI intake that triages insurance FNOL without losing adjuster control

The operational problem

First notice of loss arrived as email, web forms, and voicemail transcripts. Intake staff spent the morning reconstructing what happened instead of assigning work.

Incomplete files sat for days. Cycle time slipped. The desk hired more coordinators instead of fixing the intake shape.

What was breaking down

• Missing policy numbers and loss locations on a third of notices
• Manual assignment by whoever was least busy
• No consistent severity language for the same event type

How Striders Tech approached it

We designed intake around what an adjuster actually needs on day one, then used models to extract, score, and route — with a human confirm step on anything below confidence.

What we built

• Structured FNOL capture from email and web
• Confidence-scored extraction for policy, location, and loss type
• Routing rules by line of business and severity
• Adjuster workbench with a gap checklist, not a blank claim

Results the team can measure

• 42% drop in incomplete files reaching adjusters
• Same morning assignment on standard auto and property FNOL
• Desk capacity held steady through a seasonal spike without extra hires

Who this case study is for

Claims leaders who know intake quality decides cycle time, and who will not hand the file to a black box.

Northstar Claims operates in insurance. The work is a custom system designed around their process, not a generic template with extra fields bolted on.