Closing the subrogation gap: turning recoverable exposure into actual recoveries

Ask any claims executive whether subrogation is a priority and the answer is yes. Ask them what percent of recoverable exposure their carrier actually recovers and the answer comes with a long pause. The honest number across the industry is somewhere between thirty and fifty percent of the identifiable potential, and the gap is almost entirely workflow, not detection.

The gap is not detection

Most carriers can identify subrogation potential reasonably well. The intake form has a third-party box. Adjusters are trained to look. The mature carriers have rules engines and prior-claim cross-references. The detection layer is not the weak point.

The weak point is what happens after detection. A file gets flagged. The examiner has fifty other files in their queue. Pursuing the recovery requires assembling evidence, locating the third party, drafting the demand letter, and tracking the response — all on the schedule of the statute of limitations clock. The math, file by file, often does not justify the effort. So the file stays flagged but unpursued.

Where the recovery leakage lives

When you instrument a representative subrogation file from flag to demand-out, the same five leakage points repeat across carriers.

1. Liability evidence is not assembled inside the file. The photos, statements, and incident reports that establish third-party liability live in the file, but nobody has organized them for a recovery purpose. The examiner deprioritizes the file because the evidence work is in front of them.
2. Third-party identification stalls. The vehicle is identified but the carrier of the third party is not, or the insured did not capture contact information, and the examiner has to dig.
3. Demand timing slips against the statute clock. Statutes vary by jurisdiction and by cause of action. Files that get to demand on day 60 do better than files that get to demand on day 240. Almost nothing in the current workflow drives demand timing.
4. Demand quality is uneven. Some demand letters are forensic and well-supported. Some are template. Recovery rates correlate strongly with demand quality. Quality correlates strongly with examiner workload.
5. Response handling is reactive. When the third-party carrier responds, the file moves at the speed of the examiner's queue. Negotiation windows close.

What agents do on every flagged file

A subrogation agent operates on the entire flagged population, not the sample your examiners can reach. It does the assembly work that converts a flag into a recovery-ready file. The work is concrete.

• Reads the full file and identifies the third party with as much specificity as the evidence supports.
• Assembles liability evidence into a structured recovery packet — photos, statements, police reports, scene diagrams, and prior claims.
• Maps the statute of limitations clock per jurisdiction and per cause of action and surfaces the deadline at the file level.
• Drafts the demand letter, fully cited, on a schedule the examiner approves.
• Tracks the third-party carrier response, prepares counter-offer analysis, and updates reserve recommendations as the file develops.

The examiner remains the decision-maker on every step. The agent removes the assembly burden that previously turned a recovery opportunity into a deprioritized file.

What changes in the numbers

Subrogation is one of the few claims initiatives where the dollar lift is both large and easy to measure. The gross recovery number is in the financial statement. The closure cycle on demand-to-payment is in the claims system. The carriers we work with see the same three shifts after a full quarter of deployment.

Where to start

Subrogation deployment works best when it starts with a single liability profile inside a single LOB. Auto bodily injury is the canonical starting point for personal lines. Commercial premises liability is the canonical starting point for commercial. The reason is the same: the liability theories are well-understood, the evidence is standardized, and the statute clocks are clean to map.

1. Pick the highest-volume liability profile in your largest LOB.
2. Deploy the agent in shadow mode for two weeks. Compare its flag-to-packet output against the examiner queue's current output.
3. Move to production. Have examiners review and send demand letters on agent-prepared packets.
4. Track demand-out cycle time and recovery capture against a pre-deployment cohort.
5. Expand by liability profile, then by LOB.

The organizational change is smaller than people expect

Subrogation deployments rarely require new headcount, new vendors, or a reorg. The work that was previously done unevenly by examiners, on the files they had time for, is now done consistently by agents on every flagged file. The examiner's role becomes review and approval. The pipeline volume goes up. The dollar lift follows.

Why now

Loss ratio pressure is real in personal lines and tightening in commercial. The available levers are limited. Subrogation is the rare one that does not require a rate change, a re-underwriting cycle, or a renewal renegotiation. It requires capturing the recoveries you have already earned and have not yet collected. That is a workflow problem, and workflow problems are the problems agents are built to solve.