Building Governance Around Real AI Use

01 · Where the Engagement Started

AI had already arrived across the business. Nobody had been asked to own it.

By the time INTAI was engaged, several people across the broker's business were already using generative AI tools independently, for client emails, quote summaries, LinkedIn posts and blog drafts. Each person had reached for a tool that solved their own immediate workload problem. None of it had been coordinated, and none of it had been reviewed against a shared standard, because no shared standard existed. There was no inventory of what tools were actually in use, no policy on what could or couldn't be drafted with AI, and no distinction between content that was purely promotional in tone and content that referenced a specific product, price or recommendation.

This is not unusual, and it is not a sign of a poorly run business. The same Bank of England and FCA survey found that 84% of UK financial services firms already report having a named accountable person for their AI framework, but that figure describes what sits on paper at the level of the wider industry, not what happens function by function inside a single firm. At this broker, as at many similarly sized intermediaries, "governance" existed in the abstract long before it existed in the marketing team's actual workflow.

Leadership's own read of the situation was that this was primarily a skills and marketing gap, "we don't really understand what AI means for how we're found online." That framing was not wrong. It simply undersold the second problem sitting directly underneath it.

03 · Marketing AI Is Still Financial Promotion

Nothing about using AI to draft content changes what a regulator expects a financial promotion to look like.

The FCA's financial promotion rules require content to be fair, clear and not misleading, and that standard applies regardless of how the content was produced. The regulator does not treat this as a theoretical risk: in 2024, FCA intervention led to 19,766 financial promotions being amended or withdrawn, a 97.5% increase on the year before.

In March 2026, the FCA published its "Consumer understanding: good practice and areas for improvement" report, setting out in practical terms what it expects from firms' promotions governance: evidence-based approval steps, documented sign-off with a clear record of how wording was agreed, defined senior responsibility, and, critically, a feedback loop connecting monitoring insight back into future decisions. The report's account of poor practice reads almost like a description of ungoverned AI-assisted marketing: promotions approved without monitoring outcomes, accountability that looked clear on an org chart but wasn't clear in practice, and records that didn't show what had changed, why, or what impact it had.

None of this is AI-specific regulation. The FCA has said repeatedly that it does not intend to introduce AI-specific rules, relying instead on existing frameworks, Consumer Duty, the Senior Managers and Certification Regime, and the Systems and Controls sourcebook, to cover AI use. If anything, that makes the standard easier to miss by accident: it means AI-drafted content already sits inside the same accountability regime as everything else a broker publishes, whether or not anyone inside the business has framed it that way.

For this broker, the individual, ungoverned use of AI in marketing wasn't a small workflow shortcut sitting outside the compliance perimeter. In regulatory terms, it was indistinguishable from any other unreviewed financial promotion.

04 · Mapping Before Governing

Before building anything, INTAI mapped what was actually happening, not what a policy would later say should be happening.

The engagement began with discovery across marketing and the wider business: not the "approved tool list," which barely existed, but what staff were actually using day to day to draft, summarise and publish. That mapping exercise surfaced tools and use cases the broker's own leadership hadn't been aware of, which is typical; governance gaps are rarely visible from the top of an organisation until someone goes looking for them at the level of individual workflows.

Each use case was then classified by regulatory sensitivity, separating purely internal drafting aids from anything client-facing or touching policy wording, pricing or recommendation language. On that basis, INTAI worked with the broker to establish named ownership and clear decision rights: who could approve a new AI tool for use, who needed to sign off content before it was published, and what triggered escalation to compliance rather than routine marketing approval.

In parallel, INTAI worked with the marketing team on the search question properly, not "how do we rank on page one," but how to build consistent, accurate, well-structured information about the business, its specialisms and its people, so that being represented in an AI-generated answer meant being represented correctly. The governance framework INTAI applied drew on the structure of ISO/IEC 42001 and the NIST AI Risk Management Framework, mapped specifically against the FCA requirements, Consumer Duty, SM&CR and SYSC, that already applied to this broker's marketing and client communications.

06 · What This Means for Other Brokers

The lessons from this engagement generalise well beyond one firm's marketing team.

Ask who owns this, not just who uses it. With 84% of firms reporting a named accountable person for AI on paper but far fewer able to demonstrate it in a specific workflow, ownership needs to be tested at the level of an actual piece of content, not assumed at the level of an org chart.

Treat AI-drafted marketing content exactly like any other financial promotion. The FCA's requirement that content be fair, clear and not misleading does not change because a tool rather than a person drafted the first version, and with 19,766 promotions amended or withdrawn in 2024 alone, the regulator is actively testing that standard.

Don't confuse ranking with representation. With over two-thirds of Google searches now ending without a click, and AI Mode query volumes more than doubling each quarter, being accurately described by an AI system matters as much as, and increasingly more than, ranking for a keyword.

Build one accurate source of truth, not a pile of content. AI systems, human readers and compliance reviewers all need to be working from the same consistent facts about the business; inconsistency is the thing that breaks all three at once.

Revisit the governance whenever the tools or the rules change. The FCA's March 2026 report and its wider AI work are ongoing rather than settled; a governance model built for today's tools and today's guidance needs a scheduled reason to be checked again, not just a one-off sign-off.

References & Sources

Research and regulatory context informing this case study.

This case study draws on regulator publications, official survey data and established AI governance frameworks. Statistics are attributed to their originating source wherever directly cited.

This case study reflects INTAI's engagement approach with wealth and specialist insurance intermediaries adopting AI. Client details, including business size, sector focus and specific circumstances, have been anonymised and adapted to protect confidentiality, and any figures describing this specific engagement are illustrative rather than reported client metrics. External statistics, survey findings and regulatory publications are attributed to their original source and reflect the reporting periods of those source publications. This case study is provided for general informational purposes and does not constitute regulatory or legal advice.