More than 80% of managing general agents are already using AI, but fewer than half have a formal governance framework in place to manage it - a gap that carries consequences beyond internal operational risk. A white paper published by cyber security and IT services provider Intersys at the MGAA Annual Conference puts that figure at 52% of MGAs with governance in place, against 93% of Lloyd's managing agents surveyed by the Lloyd's Market Association earlier this year. For MGAs operating as Lloyd's coverholders rather than managing agents, that gap is not just a matter of internal readiness: managing agents with governance frameworks in place face a third-party AI risk question under their own frameworks when the coverholders below them in the capacity chain do not. The governance gap between the two tiers of the delegated authority market has a direct commercial implication for coverholder relationships and, potentially, for delegated authority reviews.
The Intersys white paper - The Perils and the Potential of AI in the MGA Market: Your Guide to Compliance-Ready Artificial Intelligence - argues that AI adoption has outpaced governance at an organisational level as well as a market one. Generative AI tools reached mainstream adoption in weeks, compared with the years it took technologies such as Netflix and Facebook to achieve comparable penetration, leaving many organisations without the governance, policies and controls needed to manage their use effectively.
Without appropriate frameworks, MGAs face specific operational and liability risks: AI hallucinations and inaccurate outputs feeding into underwriting decisions, data poisoning compromising decision-making, intellectual property loss and copyright exposure, AI-enabled cyber attacks and data leakage, regulatory failures, and heightened accountability risk for senior management in a regulated market. MGAs operate under delegated authority, making underwriting and claims decisions on behalf of insurers while handling large volumes of sensitive and personal data - which means poor AI governance has implications for regulatory compliance, capacity partner confidence and reputation, not just internal operations.
The white paper lands as AI governance draws closer FCA scrutiny. MGAs now number more than 300 in the UK, collectively underwriting over 10% of the country's £47 billion general insurance premiums, with delegated authority's overall market share forecast to exceed 45% by 2027. That growth has coincided with the regulator signalling closer oversight of delegated arrangements, particularly around claims handling, transparency and accountability between insurers, MGAs and third-party providers.
On AI specifically, the FCA has said it does not plan to introduce AI-specific rules, preferring to rely on existing frameworks including the Consumer Duty and the Senior Managers and Certification Regime to hold firms accountable. Its first Insurance Regulatory Priorities Report, published in February, named AI adoption a formal area of focus for 2026. A House of Commons Treasury Committee inquiry reported in January that despite roughly three-quarters of UK financial services firms already using AI, oversight had not kept pace, and recommended the FCA publish practical guidance on AI accountability by year end.
The April LMA AI Adoption Toolkit, produced with Barnett Waddingham following a survey covering over 60% of Lloyd's market stamp capacity, found 93% of managing agents had a formal AI governance framework in place or in development, with data privacy, cybersecurity and third-party risk among their top AI concerns. That figure now sets a de facto benchmark for the delegated authority market - and the 41-percentage-point gap between managing agent and MGA governance rates suggests the wider MGA market has significant ground to close, both to match standards expected across the broader market and to satisfy the third-party risk expectations of the managing agents on whose paper many of them write.
