Carla Gillham, Executive Head of TRUM Africa, warns that unchecked AI isn’t just an IT issue, it’s a hidden management liability that brokers need to catch before the next renewal.
Assessing a client’s true risk profile now requires digging into their digital guardrails. I’ve outlined why unchecked technological tools are quietly creating hidden blind spots in insurance programmes, along with the questions you should be asking before your client’s next renewal.
The reality is simple: AI isn’t a future risk. It’s already being used by clients for:
- Communications
- HR decisions
- Internal analysis
- Underwriting and claims support
The issue isn’t AI. It’s using AI without clear rules, oversight or accountability, and that’s when:
- Public AI tools get used with sensitive data
- AI outputs are trusted too quickly
- Decisions lose an audit trail
- Boards only see the risk after it crystallises
Where risk becomes insurable risk
From an insurance perspective, this matters. Because poor AI governance can trigger:
- Data breaches
- Misleading communications
- Biased or unfair outcomes
- Governance failures
This is no longer just an IT issue. It’s a management and liability exposure.
Questions to ask before renewal
It’s crucial to ask your clients about their AI governance now, because insurers are already looking for the same information. When reviewing their insurance, find out:
- Where AI is being used across the business, and if their data use is governed
- Whether public AI tools are restricted and how those decisions are controlled
- Who holds ultimate accountability for AI-assisted outputs
- If AI risk is actually on the Board’s radar
Unclear answers = early warning signs.
Poor governance doesn’t always remove cover, but it does create friction when a claim arises and if left unchecked, could create aggregation exposures and result in cover being excluded.
Understanding the risk is just as important as placing the cover.
AI is already influencing client decisions, often without governance keeping pace.
For brokers, the question isn’t “is AI being used?” but rather “who is accountable when it goes wrong?”
AI isn’t a risk. Unsupervised AI is. And governance gaps have a habit of appearing when the policy is tested, not when it’s placed.
TRUM Takes: helping you catch the exposures your clients don’t even know they have.
Please use this link to join the TRUM Takes series: https://trum.insure/get-trum-takes/
