The world’s largest insurance brokers recently found themselves caught in an “AI scare trade” after news that Tuio home insurance would soon be available via ChatGPT. Shares fell sharply, as investors questioned whether intermediaries could be bypassed altogether.
That reaction misunderstands where value sits in broking — and how AI is actually being deployed inside the industry. The fear was misplaced because it ignored a simple reality: broking value is rooted in judgment, structuring expertise, and market navigation — not in the mechanics of data processing
AI is not disintermediating brokers. It is changing how broking operations run.
“Disintermediation fears have followed brokers for decades. Yet in reinsurance, commercial, and specialty markets, the value of advice, structuring, and market access has only become more important over time. AI does not change that.”
Rather than replacing the role brokers play in advising on, structuring, and placing complex risk, AI is strengthening it. The firms seeing the most impact are not using AI as a bolt-on productivity tool. They are embedding it into the operating layer of the business — across placement, servicing, and renewal.
Large language models are already delivering real value across insurance workflows. They can interpret unstructured submissions, extract relevant data, and accelerate how information moves through the placement process. In that sense, they are becoming a powerful tool in the broker’s toolkit.
Increasingly, firms are recognizing that general purpose models are only part of the picture. Insurance workflows depend on domain context — policy structures, market conventions, claims histories, and regulatory nuance — that sits outside the training of generic models.
This is driving a shift toward domain-specific AI, where models are applied within the operating context of insurance itself. For example, a general purpose model may summarise a slip, but a domain-specific model can flag a missing aggregation clause or inconsistent limit structure — issues that materially affect coverage quality
The difference is material: from generating plausible outputs to delivering results that can be relied on in live placement and servicing workflows.
But their role is distinct. Brokers operate on judgment, accountability, and outcome-driven advice. In complex commercial risks, structuring and placing coverage requires an understanding of client intent, market dynamics, and the nuances of coverage that extend beyond what any model can infer from data alone.
Just as importantly, brokers bring trusted relationships across underwriting, claims, and risk management communities — applied in real time to secure the right outcome for the client.
The combination of AI capability and broker expertise is where the real shift is taking place.
That shift is most visible in how work actually flows through a broking organization.
Single Operational View of Placement Lifecycle
Historically, placement has been managed across fragmented systems — emails, spreadsheets, and disconnected tools — leaving limited visibility into where submissions stand, where delays are building, or where intervention is required. AI is changing that by enabling a single operational view of the placement lifecycle, from intake through market engagement, bind, and ongoing servicing.
For the first time, teams can see in real time what is progressing, what is stalled, and what needs attention. One of the clearest examples is in how risks are prepared before they ever reach the market. AI can now identify incomplete submissions, conflicting data, or missing documentation at intake, and guide brokers to resolve those issues before approaching insurers.
The result is a more placement-ready submission — cleaner, more complete, and more likely to attract competitive responses. That reduces rework, avoids unnecessary market churn, and improves both speed and quality of placement. Across some early deployments, teams are seeing 30%–50% fewer back‑and‑forth queries from carriers simply because submissions are cleaner before they ever hit the market.
Providing Discipline and Visibility
AI is also reshaping how brokers engage with the market itself. Rather than manually coordinating submissions and tracking responses across multiple carriers, brokers can operate within structured placement workflows — identifying appropriate markets, tracking quotes and declines in real time, and comparing terms in a consistent way. This introduces a level of discipline and visibility that has historically been difficult to achieve, particularly in complex or syndicated placements.
Crucially, the impact does not stop at bind.
Much of the operational burden in broking sits in what happens next — endorsements, renewals, claims coordination, and finance-related workflows. These have traditionally been managed in separate silos, creating friction, delays, and inconsistent service. AI is beginning to bring these activities into the same structured operating model as placement, improving responsiveness and reducing the operational drag that often erodes margins after the deal is done.
Embedding AI into the operating layer also strengthens governance. Structured workflows create clearer audit trails, more consistent documentation, and a level of process transparency that regulators increasingly expect.
The result is not simply efficiency. It is control.
And with that control comes capacity.
By reducing rework, automating routine tasks, and improving visibility into workflow bottlenecks, brokers can handle significantly higher volumes of submissions and servicing activity without a corresponding increase in headcount. In a softer market where organic growth is harder to come by, that operational leverage becomes a direct driver of margin expansion.
This is where the conversation around AI needs to move.
The dividing line is no longer between firms that are experimenting with AI and those that are not. It is between those embedding AI into the operating layer of the business and those still treating it as a series of isolated tools. The latter struggle to move beyond pilots because the underlying workflows remain fragmented. The former are redesigning how work flows across placement, servicing, and renewal — and seeing measurable gains as a result.
That is why the narrative has shifted so quickly from fear to urgency.
At Davos earlier this year, Marsh CEO John Doyle described AI as a major opportunity — not just to improve productivity, but to unlock growth and address the emerging risks facing clients. His peers across the broking sector have expressed similar views.
They are right to be confident.
Disintermediation fears have followed brokers for decades. Yet in reinsurance, commercial, and specialty markets, the value of advice, structuring, and market access has only become more important over time.
AI does not change that.
It changes how effectively that value can be delivered.
The firms that recognize this — and redesign their operations accordingly — will not be displaced by AI. The brokers who build AI-driven operating models won’t just define the next phase of competition — they will set the performance baseline the rest of the market will be forced to follow.
Related:
Topics
InsurTech
Data Driven
Artificial Intelligence
Agencies
