Intersys: AI Is Changing the Rules. Governance Must Keep Up

By Mark Kirby, Director of Professional Services, Intersys

Artificial intelligence has arrived in the MGA market with remarkable speed. From underwriting and claims to document handling and customer communications, generative AI is already transforming the way many businesses operate. For a sector that has always embraced innovation, the appeal is obvious: greater efficiency, faster decision-making, and the ability to eliminate time-consuming administrative tasks.

But there is a misconception at the heart of many AI conversations.

Generative AI is often spoken about as though it were simply another software tool. It isn’t. Large language models generate text, summaries, code and analysis by predicting the most likely response based on patterns learned from enormous datasets. They do not understand context, intent or consequences. They cannot exercise judgement or validate whether the information they produce is correct. That is why AI should be viewed less as a new technology and more as a risk multiplier if it is not properly governed.

Unlike previous technological innovations, AI is also general-purpose. It is not confined to one department or one business function. Underwriters are using it to review submissions. Claims teams are improving correspondence. Marketing teams are drafting content. Operations teams are analysing documents. AI is becoming embedded across organisations, often without a coordinated strategy, significantly expanding the potential risk surface.

That creates a challenge for every MGA.

Research carried out by Intersys in its recent White Paper – AI Governance Now – found that more than 80% of MGAs are already using AI, yet almost half have no formal AI governance framework in place. The technology is being adopted faster than organisations are putting the controls around it.

This deployment and governance gap matters because the greatest AI risks are rarely technical.

Analysis of major AI failures has repeatedly shown that the root causes are weak governance, unclear ownership and misplaced trust rather than failures in the technology itself. AI simply magnifies existing weaknesses. If an organisation has inconsistent processes, poor data quality or limited oversight, AI accelerates those problems instead of solving them.

For MGAs, the implications are even more significant.

Operating under delegated authority means making underwriting and claims decisions on behalf of capacity providers. That brings higher expectations around accountability, transparency and governance. The information passing through AI systems may include proprietary underwriting models, broker relationships, claims data, customer information and commercially sensitive intellectual property. Without appropriate controls, those assets become increasingly vulnerable.

It is also important to remember that accountability does not transfer to an algorithm.

As AI becomes more influential in underwriting, pricing and claims, senior managers remain responsible for ensuring decisions are appropriately governed. Regulation continues to evolve, but firms should not wait for prescriptive rules before taking action. Demonstrating clear oversight, documented responsibilities and evidence of governance is likely to become increasingly important as regulatory expectations develop.

Fortunately, good AI governance is neither complex nor restrictive.

The white paper identifies four principles that should underpin every MGA’s approach: maintain meaningful human oversight, ensure transparency and accountability, protect privacy and data, and actively guard against bias and discrimination. These principles recognise that AI should support professional judgement, not replace it.

Turning those principles into day-to-day practice is where governance becomes tangible. The paper’s 10 Commandments for Safe AI Use recommend practical measures such as approving AI platforms before they are deployed, providing mandatory staff training, using enterprise rather than free AI accounts, maintaining records of AI usage, controlling access to sensitive information, managing third-party risks and reviewing governance regularly. None of these measures inhibit innovation. They simply ensure AI becomes another well-managed business capability rather than an unmanaged business risk.

There is also a commercial dimension that should not be overlooked.

Capacity providers are becoming increasingly selective about the businesses they support. Strong underwriting expertise will always matter, but operational maturity is becoming an increasingly important differentiator. Firms that can demonstrate disciplined governance around AI are likely to inspire greater confidence among insurers, brokers and regulators than those relying on informal or inconsistent approaches.

The conversation around AI is often dominated by productivity. Yet productivity alone will not determine which MGAs succeed over the next decade.

The winners will be those that combine innovation with discipline. AI has enormous potential to improve underwriting, claims and operational performance, but only when it is deployed within a framework that is transparent, accountable and governed. In a market built on trust, good AI governance is no longer simply about compliance. It is becoming a competitive advantage.

Read the full paper here: AI Governance for MGAs | An Intersys White Paper