The Chief Underwriting Officer of a top-ten North American P&C carrier told me at an industry roundtable in February: "We have ninety-three AI initiatives across the company. Maybe six are in production. We have spent more on AI consultants than on actual AI engineers. And our biggest competitor just announced that their claims-triage agent is handling 40 percent of first-notice-of-loss volume autonomously. We are losing this race."
She is not exaggerating. The race she is describing is real, the gap is structural, and the agentic shift in 2026 is where the race is being decided.
The global banking industry has invested over $20 billion in AI since 2019. JPMorgan alone has more than 2,000 AI-related roles. Goldman Sachs processes millions of data points through ML models daily. Banks are not experimenting with AI anymore. They are operating with it.
The insurance industry is eight to ten years behind. According to McKinsey's 2024 Global Insurance AI Survey, fewer than 10 percent of P&C and life insurers have deployed AI at scale, compared to over 30 percent of large banks. Deloitte's 2024 insurance industry outlook found 73 percent of insurance executives acknowledge their AI capabilities are "nascent or developing." Accenture's 2025 insurance AI value study estimates the industry is leaving $170 billion in annual value on the table by failing to adopt AI and advanced analytics.
The gap is not about ambition. Every insurance CIO we speak with has an AI roadmap. The gap is about architecture and the specific structural barriers that make insurance AI harder to operationalize than banking AI. And it is about to be tested by the agentic transition that is already running in banks. Lemonade and Ping An's published deployments are early signals of what is coming for the rest of the industry.
Why insurance is structurally different
Banking and insurance share regulatory complexity, data volume, and risk management requirements. But insurance carries four architectural barriers banking does not.
Unlock the Future
Continue reading in the FinScale Magazine
This insight was originally published in the third issue of FinScale Magazine by TrialScale. Download the magazine to keep reading.