The Revolution: AI in Underwriting

Could Delegated Authorities Become Redundant at Lloyd's?

Summary

• Delegated authority business now exceeds 40% of Lloyd's premium income — a model that doubled in volume since 2018 to £22.1bn, yet carries escalating governance, audit and oversight costs.

• Agentic AI — multi-agent systems capable of end-to-end workflow execution — is transitioning from pilots to production in 2026, creating a credible alternative to the traditional coverholder model.

• AI agents embedded in syndicate core systems can underwrite, price and bind risks directly, eliminating the need for delegated authority and replacing it with AI-augmented line slips.

• The reformed model repositions coverholders as brokers, streamlines capacity deployment, removes intermediary overhead and sharply reduces adverse selection and moral hazard.

• Speed, scalability and cost reduction are the headline gains; regulatory explainability, data governance and market cultural resistance are the primary risks to navigate.

• Early-mover syndicates in cyber, property, SME and niche lines are best positioned to pilot the transition — with broader market adoption expected by 2027.

In the heart of the London insurance market, delegated authorities - binding authorities - granted to coverholders - have long been a cornerstone of Lloyd's operations. These arrangements allow syndicates to extend underwriting capacity to third-party coverholders/managing general agents (MGAs), who bind risks, issue policies and often handle claims on behalf of syndicates. By 2023, delegated authority business accounted for more than 40% of Lloyd's premium income, with volumes doubling from £10.4 billion in 2018 to £22.1 billion. Coverholders, numbering in the thousands globally, provide local expertise, rapid distribution, and access to niche or international markets that syndicates could not serve efficiently from the Room alone.

The Enduring Power — and Hidden Costs of Delegated Authority

Yet this model, while successful, carries persistent friction. Oversight is intensive: Syndicates must conduct regular audits, enforce reporting standards via bordereaux, monitor performance metrics, and manage third-party material issues ranging from data delays to poor risk selection. Lloyd’s has kept a sharp eye on these facilities, pushing hard on governance to stop underperformance from dragging down overall results. The administrative burden—audits, compliance checks, reconciliation of premiums and claims processing—adds real cost and delay, while exposing the market to conduct, credit and reputational risks.

Enter Agentic AI: A Game-Changing Force

Enter AI agents: autonomous, intelligent systems capable of operating directly within an underwriter's digital environment. As of 2026, agentic AI—multi-agent ecosystems that reason, decide and execute end-to-end workflows—is moving from pilots to production scale. Insurers are already deploying it for real-time risk assessment, dynamic pricing, document ingestion, fraud detection and policy binding. Projections suggest AI agents could handle underwriting submissions, escalate only complex exceptions to humans, and cut processing times by 40-70%. In this context, the traditional delegated authority model starts to look not just cumbersome but potentially obsolete.

A Transformative Proposition: Direct AI Underwriting via Line Slips

The proposition is straightforward and transformative: embed AI agents inside the syndicate's core systems to underwrite risks directly. No delegation required. Coverholders, freed from binding authority, transition into pure brokers who source and introduce business. Placement shifts to line slips—a long-established Lloyd's mechanism that already streamlines multi-syndicate participation.

A line slip is an agreement under which the lead syndicate is authorised to bind risks on behalf of the following syndicates. Unlike binding authorities, line slips do not grant or delegate underwriting power to intermediaries; the lead (in this future scenario, an AI-augmented underwriter) quotes and binds within predefined parameters. Followers adhere automatically, with risks bulked via bordereaux or processed digitally. Documentation follows Market Reform Contract (MRC) standards or as may be agreed, ensuring speed and consistency. Syndicates require no separate Lloyd's approval to operate line slips, making them agile tools for capacity deployment.

How the New AI-Driven Underwriting Workflow Would Actually Work

In an AI-driven workflow, a broker (the former coverholder) submits a risk via a digital platform for insurance underwriting. The AI agent receives data that is enhanced in its accuracy - trained on vast datasets of historical claims, real-time telematics, satellite imagery, IoT signals and market pricing - assesses it instantly. It applies the syndicate's risk appetite, calculates premiums dynamically, and binds the risk if within parameters. The AI lead syndicate underwriter triggers the line slip, distributing participation across syndicates seamlessly. No human underwriter needs to review routine cases; exceptions route to specialists. The entire insurance underwriting process happens in minutes, not days or weeks.

The End of the Tedious Rigmarole: Goodbye to Traditional Oversight

This model eliminates the tedious rigmarole of oversight and audit that currently burdens delegated authority arrangements. No more third-party audits, delayed bordereaux or reconciliation disputes. Data flows directly from the AI system - structured, auditable and real-time - into syndicate platforms. Regulatory reporting becomes automated. The former coverholder, now acting purely as a broker, focuses on client relationships, risk consulting and distribution without the compliance overhead of delegated underwriting. Their role aligns more closely with traditional Lloyd's brokers: placing business with capacity providers who retain full control.

Compelling Artificial Intelligence Benefits for the Market

The benefits are compelling. Speed and efficiency soar; insurance underwriting using enhanced data, instant pricing and binding could unlock new distribution channels, such as embedded insurance in digital platforms. Accuracy improves through data-driven decisions with AI underwriting, reducing adverse selection and loss ratios. Costs plummet: Syndicates save on audit programmes, oversight teams and intermediary margins tied to binding authority. Scalability across the market explodes - AI agents handle volume without proportional headcount growth. Moral hazard diminishes because the underwriter (using AI driven underwriting) retains direct authority, aligning incentives perfectly with syndicate capital providers.

Positioning London for Competitive Advantage with AI Solutions

Lloyd's and the broader London market stand to gain competitively. The market has long prided itself on innovation, from the Lloyd's Lab incubators to earlier digital initiatives. AI agents could accelerate this, positioning London ahead of continental or US competitors still reliant on legacy MGAs. Smaller or specialist risks, currently underserved because of delegation economics, become viable. Capital efficiency rises as syndicates deploy capacity more precisely through line-slip facilities.

The Challenges That Cannot Be Ignored - Regulatory, Cultural, and Technical Hurdles

Yet challenges loom. Regulatory compliance hurdles are significant. While Lloyd's and the FCA/PRA encourage innovation, AI decision-making must meet standards for fairness, explainability and accountability. Algorithmic bias, data privacy (especially under GDPR and emerging AI rules) and cyber resilience would require robust governance frameworks. Insurers cannot simply 'set and forget' autonomous agents; human oversight for high-stakes or novel risks remains essential, at least in the transition phase.

Market acceptance poses another barrier. Underwriters and brokers steeped in relationship-driven culture may resist ceding judgment to machines. Syndicates must invest in integration—modern, cloud-native core systems capable of hosting agentic AI—while legacy platforms lag. Talent shifts will occur: traditional underwriters evolve into AI supervisors and exception handlers, while data scientists and prompt engineers become central. Job displacement fears could spark resistance, though history shows insurance adapts (as it did with computers and early analytics).

Data Quality, Governance and Ethical Considerations

Data quality and model governance are critical. AI agents require clean, comprehensive inputs: poor data in yields poor underwriting out. Lloyd's efforts to standardise data already push in this direction—Artificial intelligence would amplify the need. Ethical considerations arise: who is liable if an AI misprices a catastrophic risk? Contracts and professional indemnity policies must evolve to address AI errors.

A Phased Transition: What the Road Ahead Looks Like

Transitioning would not happen overnight. Hybrid AI models are likely first - AI-augmented coverholders feeding data to syndicate AI leads, or pilot line slips with AI underwriting desks. Early adopters, perhaps specialist syndicates in cyber, property, SME or niche lines, could prove the concept. Some carriers are already experimenting with AI in binding authority workflows; scaling this internally to eliminate delegation is the logical next step.

AI in Insurance: The Human Element Still Matters

By 2026-2027, as agentic AI matures, the economics may become irresistible. Delegated authority business, projected to exceed 45% of Lloyd's premiums by 2027 under current trends, could plateau or contract as direct AI underwriting via line slips proves superior. Coverholders rebrand as brokers, focusing on value-added services. The London market streamlines: fewer intermediaries in the chain, faster capital deployment and stronger alignment between risk originators and capital providers.

This is not the end of human expertise—far from it. Complex, bespoke risks will still demand nuanced judgment, negotiation and relationships. AI agents augment rather than replace; they handle the volume and routine, freeing talent for high-value work. Lloyd's, with its 300+ year history of adaptation—from coffee-house origins to digital transformation—stands poised to lead once again.

Embracing the Inevitable: Lloyd’s Next Chapter

The question is not whether AI will disrupt delegated authorities, but how quickly the market embraces the shift. Those who integrate AI agents into their underwriting core, leverage line slips for efficient capacity and reimagine former coverholders as brokers will thrive. The rest risk being left with the overhead of yesterday's model in tomorrow's market. The London insurance market's future is not delegated - it is direct, intelligent and instantaneous.

Next Steps for Transitioning to AI in Insurance Underwriting

• Commission an internal feasibility study: map your Delegated Authority book by risk class and identify where AI-direct underwriting is technically ready today.

• Engage your technology leadership to audit core system readiness for agentic AI integration — cloud-native capability is a prerequisite.

• Initiate dialogue with Lloyd's Delegated Authorities team and the Lab to explore regulated pilot structures for AI-led line slips.

• Reframe coverholder relationships now: begin transitional conversations with key partners about the broker-oriented model before the market shifts beneath them.

• Appoint a dedicated AI Underwriting Transformation lead with joint accountability across underwriting, technology and compliance.