Most insurance companies that outsource operations would not renew those contracts on current terms, according to new research. The biggest complaint is that providers are falling short on AI.
Only 8% of buyers of outsourced insurance operations said they would renew their contracts as they stand, according to The State of Insurance Operations, published by technology research and advisory firm ISG. Some 64% said they would renegotiate for AI-driven efficiency, and 23% said they would rework terms and service levels entirely.
The study surveyed 150 insurance decision-makers in August 2026, across carriers, brokers, managing general agents (MGAs) and reinsurers in North America, Europe and Asia. It covered 20 operational activities. The research was commissioned by mea Platform, which sells AI software for insurance operations.
AI-led innovation was the area where outsourcing providers most often fell short. Some 49% of business process outsourcing (BPO) users said it was below expectations, the largest shortfall among the nine areas assessed. Strategic value came next, at 39%.
“The shortfall buyers name most often is AI-led innovation, and that is a hard thing to answer at a renewal, because it cannot be settled with a price concession,” said Ashish Jhajharia, insurance principal analyst at ISG.
ISG recommended that insurers compare outsourcing providers, technology platforms and in-house builds on equivalent evidence of how each performs on real workflows.
Insurance outsourcing spans much of the value chain, including submission intake and underwriting support, policy administration and billing, claims handling from first notice of loss to subrogation, bordereaux and reinsurance accounting, and regulatory reporting. In life and annuities, third-party administrators often run entire blocks of business, including closed books, on their own systems.
Estimates of the market’s size vary widely, depending on what is counted. Market research firms cited in the release put core insurance BPO at between $7.5 billion and $8.4 billion in 2025. A broader definition including technology-enabled services puts it at $68.4 billion for 2026.
Most outsourcing is priced per transaction or per full-time employee. That means costs rise in line with volume, which is the model AI vendors argue they can break.
Buyers are cautious about how much they hand over to AI. Asked what they would trust for high-stakes underwriting and claims decisions, 75% chose an insurance-specific AI model or a governed hybrid of AI and human oversight. Only 6% would accept a general-purpose AI model on its own, and 20% would not trust AI with limited oversight at all.
The report set out four conditions for trustworthy AI-driven work. Decisions should be based on controlled evidence such as policy wording, endorsements and claims guidance. There should be clear permissions separating what a system can read from what it can decide. A named person should be able to override or stop it. And it should operate within the institution’s own rules.
"Outsourcing to cheaper labor buys a few points of unit cost, and it takes years of transition and governance to collect them," said Martin Henley (pictured), CEO of mea Platform. "What insurers can buy now is execution across that entire scope, from submission intake through policy administration, claims and reinsurance accounting, and it goes live in weeks because the system already understands insurance."
For carriers, brokers and MGAs with outsourcing contracts coming up for renewal, the findings suggest buyers now have leverage. With most unwilling to renew on existing terms, renewals are a chance to push for AI-driven efficiency, clearer service levels and pricing that doesn’t simply scale with volume.
The governance findings are just as important. Insurers and brokers remain accountable to regulators and clients for outsourced work, whoever or whatever does it. Any move towards AI-driven operations will need clear human oversight and audit trails, especially for underwriting and claims decisions that affect customers.
The commercial interest behind the research is also worth weighing. mea Platform commissioned the study and sells the alternative it points towards, so buyers should test any vendor’s claims against their own workflows, as ISG itself recommends.
