The real AI battleground in broking isn’t client-facing

As platforms embed machine learning, competitive advantage may shift behind the scenes

The real AI battleground in broking isn’t client-facing

Transformation

By Bryony Garlick

When the first ChatGPT-integrated broker app made headlines, talk of disintermediation resurfaced almost immediately. If AI can match risks, surface policy options and draft responses, what happens to the advisory role?

At least one industry professional dismissed the reaction as overblown, arguing that similar fears emerged in the legal profession when AI tools first entered workflows - yet lawyers were not replaced. 

The legal comparison is not exact, but it is instructive. When generative AI tools began entering legal workflows in 2023, headlines predicted significant disruption. Research from Thomson Reuters has since shown rapid uptake of AI in drafting, document review and research tasks. What has not followed is wholesale replacement. Instead, AI has been embedded within processes, productivity has shifted, and oversight requirements have increased, while qualified judgement remains central.

For brokers, however, the structural dynamics differ in one important respect: infrastructure.

Many UK brokers do not control their own core technology environments. They rely on broker management systems and outsourced SaaS platforms to run policy administration, workflow, trading connectivity and data capture. In the UK market, providers such as Acturis underpin significant volumes of premium across brokers, MGAs and insurers. These platforms are not peripheral tools; they form the operational backbone of day-to-day broking activity.

That dependency reframes the AI debate. If machine learning and predictive analytics are embedded at platform level, competitive advantage may depend less on individual experimentation with standalone AI tools and more on the pace and direction of vendor development.

This shift is already under way. Broker platform providers have begun integrating predictive analytics and machine learning into their systems, surfacing renewal propensity, identifying cross-sell opportunities and automating aspects of submission handling. At the same time, operational capacity remains a real constraint across the market. Industry commentary has highlighted that many carriers are unable to review 100% of submissions because of sheer volume and human resource limitations. AI-driven triage and ingestion tools are increasingly positioned as a means of reducing missed opportunities and accelerating response times.

As another industry voice put it, the urgency around AI has shifted from curiosity to priority - “an infrastructure move of survival.”

The risk for brokers, therefore, may not be sudden disintermediation but gradual divergence. Firms operating on AI-enhanced platforms could improve speed-to-quote, refine data quality and identify revenue opportunities earlier in the cycle. Those operating on slower or less adaptive systems may find themselves competing at a structural disadvantage.

Speed-to-quote influences win rates. Workflow efficiency influences cost ratios. Data quality influences underwriting appetite and placement outcomes. If AI capability becomes embedded within core systems rather than layered on top, then the competitive divide may emerge quietly but decisively.

This is where the legal analogy becomes relevant again. AI did not eliminate lawyers; it reshaped workflow and raised productivity expectations. The same may prove true in broking. The advisory function remains grounded in judgement, negotiation and regulatory accountability. But the efficiency with which brokers operate, respond and process risk may increasingly depend on the systems beneath them.

In that context, the strategic question shifts. It is no longer simply whether AI will replace brokers. It is whether brokers’ technology partners are evolving quickly enough to support their competitiveness in an AI-enabled market.

The real battleground may not be client-facing at all. It may sit inside the platforms that brokers already rely on, and in how effectively those platforms embed intelligence into everyday operations.

If so, the next phase of AI in broking will not be defined by whether advice is automated, but by whether infrastructure keeps pace with ambition.

Related Stories

Keep up with the latest news and events

Join our mailing list, it’s free!