A legacy claims engine on a sunset clock, and a process that had to land somewhere.
The carrier's existing claims engine was being decommissioned. The intake and examiner workflow that ran on top of it had years of accumulated logic: multiple entry points, rules that decided which claims needed human review, document handling, examiner queues, and downstream handoffs to a database of record. None of it could simply be turned off.
The replacement had to do more than survive the cutover. It had to give examiners a faster surface, cut the volume of claims a human ever had to open, and keep document storage costs from following the workflow into Salesforce.
The WorkTriage at the door. Automate what should be automated. Keep humans in the loop where it matters.
- Unified intake across every channelBuilt a Service Cloud intake engine with multiple entry points, including a pipeline for physical mail claims. These were processed by an internal vendor, scanned, and entered directly into the system without manual reentry. Every claim, regardless of how it arrived, lands as a structured case in one place.
- Automated routing for rules-eligible claimsStood up a criteria-driven process that identifies claims meeting defined adjudication rules, routes them into a dedicated queue, and batches them for downstream submission to the database of record. Claims that fit the pattern never need a human pair of eyes.
- Examiner workflow with flows and assignment rulesFor claims that need review, examiners can manually enter data, trigger flows, and reassign cases through native Salesforce case assignment rules. Custom UI handles multiple claims on a single case, while supporting hundreds of claim line items that can be easily used and manipulated.
- Document viewing on AWS S3, not in SalesforceBuilt a customer image view that renders claim documents through AWS S3 presigned URLs. Examiners see what they need inside the case record. Salesforce never stores the files. File-storage costs stay on the cheaper tier where they belong.
- Cleared the cutover without breaking the floorMigrated the operational logic off the legacy engine and onto Service Cloud before the platform sunsetted. No gap in claims processing, no workaround period, no inbox queue of unprocessed mail waiting for someone to find a fix.
The cheapest claim to process is the one a human never opens.
A like-for-like rebuild would have kept the carrier in business. It would not have moved the work. The win was in what the engine decides on its own: which claims auto-adjudicate, which ones a human sees, and where the documents live while all of that happens. Service Cloud carries the workflow. S3 carries the files. Examiners carry the judgment calls that actually need them.
