Artificial intelligence is becoming central to the future of property and casualty (P&C) insurance, with carriers investing heavily in tools designed to improve underwriting, portfolio oversight and operational efficiency. Yet as AI adoption accelerates, are insurers gaining better control of their portfolios, or simply adding another layer of technology to already fragmented operations?
New research from Federato suggests the answer depends less on how much insurers spend on AI and more on how effectively they integrate it. The company’s 2026 State of P&C Insurance Technology report, based on responses from 750 professionals across insurers, managing general agents (MGAs) and MGA aggregators, found a 64 percentage point gap between leadership’s perception of portfolio visibility and the experience of underwriting teams.
While 91% of executives believe they have real-time control over their portfolios, only 27% of underwriters agree. According to the report, organisations that have fully embedded AI into underwriting workflows are 3.6 times more likely to achieve genuine portfolio control than firms layering AI onto fragmented legacy systems.
The findings highlight a broader challenge facing the insurance sector. Many organisations have prioritised AI deployment, but fewer have addressed the disconnected systems, inconsistent data and fragmented workflows that limit the technology’s effectiveness. Without stronger integration, AI risks becoming another isolated tool rather than an engine for better underwriting decisions.
That disconnect is also reflected in underwriting governance. Federato found that while 93% of insurance leaders believe underwriting guidelines are applied consistently, 88% of employees said decisions regularly deviate because of missing data, disconnected systems and workflow constraints. As insurers expand products and distribution channels, maintaining consistent underwriting practices becomes increasingly difficult when critical information remains spread across multiple systems.
Operational inefficiency continues to compound the problem. The report estimates that the average P&C employee spends around five hours each week manually coordinating information across disconnected platforms, representing an annual productivity cost of more than $10,000 per employee. At the same time, severe appetite drift nearly doubled over the past year, rising from 18% to 39% despite wider AI adoption, suggesting that investment alone is not preventing execution from drifting away from strategic objectives.
Governance risks are also becoming more pronounced. According to Federato’s research, 89% of employees admitted using unsanctioned shadow AI tools, indicating that staff are increasingly turning to consumer AI applications when enterprise systems fail to support their daily work. As AI becomes more embedded across insurance operations, these unofficial workarounds could introduce additional risks around compliance, security and oversight.
For more, read the full Federato report here.
Copyright © 2026 FinTech Global
