S04.10 · Insurance
Outsourced claims handling and loss adjustment where AI-driven photo and drone estimating directly compresses the per-claim adjuster cost.
Claims, third-party administrator (TPA) and adjusting firms provide outsourced claims handling and loss adjustment for carriers and self-insured employers. No single defensible aggregate market figure exists — estimates vary widely depending on whether the scope is health TPA or P&C claims and adjusting — but core P&C independent-adjusting and TPA services run roughly $40-60B globally (2024E, IMARC, Mordor Intelligence), with broader "insurance TPA" definitions running materially higher. The uncomfortable fact about this segment is that its revenue unit and its AI exposure are the same thing: fee-per-claim economics rest on the per-claim adjuster's time, and that time is precisely what agentic AI removes.
Forward growth is cited as high-single to low-double-digit across research firms, driven by growth in self-insurance and by catastrophe-claims surges that outstrip carriers' and self-insureds' internal claims capacity. The US is the largest market given its self-insured-employer and workers'-compensation structure, and catastrophe-driven independent-adjusting demand spikes regionally — the hurricane-exposed Gulf Coast and wildfire-exposed West are the two clearest examples. A handful of national TPAs and adjusting firms, including Sedgwick, Crawford & Company and Gallagher Bassett, hold significant share, while catastrophe and independent adjusting itself remains fragmented among contract adjusters who scale up and down with event volume.
The segment sits downstream of carriers and self-insured employers and upstream of repair networks, medical bill review and legal/subrogation vendors. Economics are fee-per-claim or FTE-based, asset-light, with cyclical volume spikes tied to catastrophe events and low capital intensity — a services business, not a risk-bearing one, which is the structural reason it can flex claims capacity with event volume in a way a carrier's own staff cannot. That flexible cost structure also explains why self-insured employers outsource rather than build: a workers'-compensation or general-liability claims function sized for peak volume sits idle most of the year if run in-house. Buying capacity by the claim beats owning it by the head.
One adjacency adds a revenue line, the other protects one. Medical bill review integration is expansionary fee capture, a second stream layered onto the same claims relationship. Catastrophe surge-staffing platforms are the defensive counterpart, insulating revenue against the volatility inherent in event-driven demand.
Agentic AI automates first-notice-of-loss triage and damage estimation from photo and drone imagery, and increasingly handles claims correspondence directly. That collapses the per-claim adjuster labor cost — the core moat this entire segment has been built on, since fee-per-claim economics only work if the claim actually requires a human adjuster's time. Automation is already underway and well advanced for property and auto physical-damage claims, where photo-based estimation is mature. Complex casualty and liability claims, where litigation exposure and negotiated settlement still require human judgment, remain 5-10 years from full automation.