Githesh Ramamurthy used two sentences on CCC Intelligent Solutions' July 30 earnings call to disclose a shift that reaches past claims operations into how casualty carriers reserve: "Liberty Mutual decided to deploy a significant portion of its casualty business on the CCC platform, and Allstate selected CCC for its third-party casualty operations" (CCC Q2 2026 earnings call, July 30, 2026). Two top-five auto insurers now route bodily-injury documentation through variants of the same vendor's rules engine.

Key Takeaways

  • 225 million medical documents a year already flow through CCC's casualty unit, the raw material for Medhub, the medical-synthesis engine adapted from the EvolutionIQ workers' compensation product to auto bodily-injury demand packages.
  • Bodily injury is now 52.3% of combined auto claim dollars paid, up from 44.4% in 2022, even as physical-damage claim frequency fell 14% over the same two years.
  • Average third-party submitted medical expenses reached $32,300 in the first quarter of 2026 against $24,300 in the first quarter of 2023, a 33% rise in three years that outpaces any plausible medical-trend assumption.
  • $2 billion of commercial auto liability reserve deficiencies were added in 2025, against a 114.7 combined ratio on other liability (occurrence). The triangles absorbing this platform change are already running deficient.

What CCC Told Investors on July 30

The casualty disclosure landed inside a broader print. Total revenue reached $285.9 million in the second quarter, up 9.8%, and AI-based solutions crossed $120 million in annualized revenue, growing nearly 50% year over year and now roughly 11% of total (CCC Intelligent Solutions, July 30, 2026). Management guided full-year revenue to $1.158 billion to $1.164 billion.

Metric Figure Source
CCC total revenue, Q2 2026 $285.9M, up 9.8% YoY CCC, GlobeNewswire, Jul. 2026
CCC AI solutions revenue >$120M annualized, ~11% of revenue, up ~50% YoY CCC Q2 2026 earnings call
Liberty Mutual casualty NWP, Q2 2026 $717M, up 21.7%; pricing up 6.2% Liberty Mutual, PR Newswire, Aug. 2026
Allstate auto reserve releases, 2026 YTD $1.5B, 51% from accident years 2023-2024 Allstate Q2 2026 results
Other liability (occurrence) industry underwriting loss, 2025 ~$11B, improved from ~$13.7B in 2024 AM Best, Insurance Journal, Jul. 2026
Bodily injury share of combined auto claim dollars, 2025 52.3%, up from 44.4% in 2022 CCC Crash Course, Claims Journal, Aug. 2026

The AI figure is not new ground; the site covered CCC crossing the 10% AI-revenue threshold in the first quarter. What is new is the specificity behind the casualty line. Ramamurthy named the product: "It's a solution called MedHub, which uses a synthesis capability to synthesize and provide guidance for very complex medical claims for auto." It had been announced in April 2025 as an extension of the EvolutionIQ acquisition, adapting a workers' compensation medical-synthesis engine to auto bodily injury.

That is a different data problem than the one CCC built its business on. Auto physical damage is photo-based estimating and parts pricing, where the loss is a fixed, largely undisputed number within days. A bodily-injury file accumulates medical records, demand letters, treatment codes and attorney correspondence over months or years, and the number is negotiated rather than appraised. Bodily injury now accounts for 52.3% of combined auto claim dollars paid, up from 44.4% in 2022, while average third-party submitted medical expenses climbed to $32,300 from $24,300 in three years.

What a Shared Casualty Data Layer Does to the Triangles

Loss development triangles assume the pace and pattern of claim reporting, case reserving and payment reflect a consistent claims-handling process across accident years. A carrier migrating a significant share of its casualty book onto a new triage platform mid-stream breaks that consistency at the exact point the actuary needs it most.

If Medhub-assisted claims move from first notice of loss to a documented, synthesized medical picture in weeks rather than months, case reserves on newer accident periods will look more complete, sooner, than the historical pattern the link ratios were built on. A development factor selected off five years of manual-triage experience will systematically overstate the tail on claims running through faster automated intake, not because ultimate severity changed but because the reporting and case-reserving pace underneath it did.

The risk cuts the other way too. Faster documentation can mean a carrier recognizes case reserve adequacy, or inadequacy, sooner than its triangle predicts, pulling forward strengthening that used to surface as adverse development two or three years later. Distinguishing "the platform found the true severity faster" from "the platform produces a systematically different severity read" requires tracking Medhub-assisted claims separately from the legacy population for at least two to three accident years before blending them.

The pressure that sits underneath makes the distinction expensive to get wrong. Other liability (occurrence) posted an industry underwriting loss near $11 billion in 2025, improved from about $13.7 billion, but net losses incurred hit a five-year peak near $50 billion against a 114.7 combined ratio, with commercial auto liability adding $2 billion of reserve deficiencies. The site's earlier look at record adverse development on casualty triangles made the same point about a different input: the triangles were already absorbing a claims and litigation environment that had shifted faster than the link ratios assumed, before any platform migration entered the picture.

Two Rivals, One Vendor

Liberty Mutual and Allstate compete directly in personal auto, and both now run a meaningful share of casualty claims through the same underlying platform logic. That is a different kind of vendor concentration than an AI underwriting tool, because a casualty triage platform does not just score a risk at binding. It shapes what data gets captured, in what structure, and on what timeline, for every claim that flows through it afterward.

When two of the largest carriers standardize that intake layer on one vendor, the industry gets something closer to a common casualty data schema than it has had. That is useful for benchmarking, since claim characteristics become more directly comparable. It is also a channel through which any systematic bias in the triage logic, on comparative-negligence handling in a specific state or on how a model weighs a particular injury code, shows up correlated across multiple carriers' books at once rather than isolated to one.

CCC's customer base already spans 27 of the top 30 US auto carriers by direct written premium, a concentration the site flagged in its coverage of the top-five subrogation win. A casualty triage platform sits closer to reserve inputs than either that or the workers' compensation rollout, because it structures the medical and liability narrative an adjuster and eventually an actuary will rely on.

The documentation is also getting faster on the other side of the file. Plaintiff firms adopting generative AI for demand-letter preparation rose to 41% of surveyed firms in 2026 from 28% in 2025, so the material arriving in a carrier's claim file is increasingly AI-assembled before the carrier's own AI touches it. A triage engine on the receiving end is not automating a stable workflow; it is the first structured checkpoint applied to narratives that are themselves being standardized by a counterparty with opposite incentives.

Further Reading on actuary.info