The Claims Revolution That's Already Too Late
The $100B Opportunity in Insurance Claims
Claims are the operational core of insurance, influencing everything from underwriting decisions to customer retention. They consume 10-20% of every premium dollar and carry the highest stakes: reputation, trust, and margin are all on the line. Yet most insurers have outsourced this critical function to third-party administrators (TPAs), creating a structural disconnect. What’s more, I’ve yet to meet an insurer who’s truly happy with their TPA. These are businesses with no margin to invest in better technology or service. Take Crawford, one of the largest TPAs: $1.35B in revenue, but trading at just $550M market cap, with razor-thin profits ($3K per employee) and a 1.5-star Google rating littered with customer complaints. Despite the poor performance of TPAs, carriers are reluctant to bring claims in-house. The result is a broken model that leaves everyone dissatisfied.
This dynamic has created a massive opportunity for an AI-native TPA that can tackle mid-to-high complexity claims with both superior technology and service delivery. The timing couldn't be better for disruption in this space. Bain estimates that generative AI could eliminate 20-25% of today's loss-adjustment expenses and 30-50% of claims leakage – representing a $100 billion global opportunity in property and casualty claims alone. The question is no longer whether AI belongs in claims, but who will deploy it most effectively to capture this enormous market opportunity. As Bronek Masojada, former CEO of Hiscox, tells me: “Improving claims customer satisfaction and reducing leakage is the holy grail. The team who is able to do this will win.”
This winning strategy lies in building a new kind of TPA that combines AI with complex claims expertise. Despite representing a minority of claim volume, complex cases account for half or more of total claim payouts. This is also where the real margin exists and where carriers depend most heavily on their TPAs. A complex claim typically involves multiple factors that require judgment and expertise. Examples include:
Coverage disputes with ambiguous policy language or contested liability
Multi-party liability such as commercial auto accidents, construction defects, or product liability cases
High-value property losses that require specialized appraisers, engineers, or forensic experts
Catastrophic personal injury where long-term care assessments and life care planning are needed
Commercial property damage that require business interruption calculations, and equipment valuations
Regulatory complexity such as claims crossing state lines or involving specialized industries
Importantly, no insurer wants to split their claims between different providers for simple versus complex cases; the TPA that can handle complex claims will naturally inherit the simple ones too. While no TPA has successfully gone the fully automated AI route due to compliance concerns, lack of investment capital, and the relationship-driven nature of insurance, the opportunity exists for a digitally native player to blend human expertise with AI efficiency.
In an industry where trust is everything and outcomes are high-stakes, the next-generation TPA must be built from the ground up with both technical rigor and domain depth. The incumbents can't retrofit their way into this future, and pure-tech plays have already proven they can't navigate the compliance and relationship requirements that define complex claims. A new kind of player – AI-native, service-oriented, and battle-tested in complex claims – has the opportunity to not just improve margins and customer satisfaction, but to become the standard that redefines what carriers expect from their most critical operational partner. The winner won't just capture complex claims, they'll inherit the entire claims portfolio as carriers consolidate around a single trusted partner.