Melody Cloete | AI can assist, but it cannot advise

AI has the capability to compare products in seconds. It can summarise benefits and generate comparisons that appear logical, and many advisers are using AI tools to do just this. This is not necessarily a bad thing, but there are clear boundaries for where AI's role should start and end.
 
The truth is that, if a client's claim is rejected, AI will not sit across the table from them explaining why. This responsibility still remains with you.
 
The danger with AI in financial advice is not always obvious, but the results may be overlooking important detail. Life insurance advice is a difficult space for AI to operate in, as tools can sound convincing while missing critical nuances. Products are technical, definitions are complex, and outcomes depend on detail. Critical illness, for example, has hundreds of variations of conditions like cancer, and cardiovascular and neurological conditions, and each life insurer's product has its own wording and requirements. And, while two products can appear similar on paper, they can perform very differently when a claim is paid out.
 
A recent comparison of early cancer definitions highlights this risk. Some insurers limit cover to a narrow list of cancers. Others require specific treatments before a claim is paid, while others provide cover for any carcinoma-in-situ without linking it to treatment (which is the easiest definition to claim from). When the same set of definitions across 10 life insurers was analysed using different AI tools, the results were inconsistent. One tool prioritised broader wording and ranked that insurer's product highest. Another focused on the presence or absence of treatment requirements, while a third placed more weight on exclusions. This led to life insurers that were ranked highly by one tool while being rated much lower down by others.
 
On the surface, the individual outputs looked well-structured and reasoned but, when placed side-by-side, it became clear that the tools could not interpret the practical impact of the wording or recognise where simplicity in a definition could lead to better claim outcomes. Because AI tools are not all designed in the same way, using them to compare product definitions in the life insurance context becomes complicated and prone to inaccuracies.
 
This subjectivity makes it difficult to identify when an answer is off-track. The output looks complete. It compares multiple features. It uses clear language. It is easy to assume that it is correct. But this is where human judgement matters.
 
An experienced adviser understands that good advice is not simply about comparing wording. It is about understanding context, identifying risks that may not be immediately obvious, balancing trade-offs, and applying product knowledge to a client's specific circumstances. AI does not possess lived experience or the ability to exercise judgement under uncertainty.
 
Further, if advice based on AI is wrong, the consequences sit with you. A recommendation that does not cover a client's key risk could lead to a declined or reduced claim, and this can result in a complaint and escalation to the Ombud, where you may need to justify the basis of your advice and how you reached your conclusion. Findings of poor advice or negligence can include personal financial liability for your client's losses, and reputational impact.
 
But there are practical steps you can take to protect yourself and your clients. Treat AI outputs as a starting point, not a final answer. Validate what you receive against product material and speak to product specialists. Draw on your own experience. Most importantly, do not outsource your judgement.
 
At the same time, AI can add real value when it is used in the right way. It can help you explain concepts like the role of income protection alongside lump sum benefits, or the need for cover beyond group schemes. It can support the drafting of emails and marketing campaigns. In these areas, the risk of harm is lower and the gains in efficiency are clear.
 
In this age of AI, your role as an adviser remains unchanged. You interpret information in the context of a client's circumstances, guide decisions when there is uncertainty, and help your clients weigh options and understand trade-offs. You bring empathy that no tool can replicate. The value you provide does not sit in the information alone. It sits in how you apply it.
 
Simply put, AI can support your work and improve how you engage with clients, but it should not replace the thinking behind your recommendations.

*Melody Cloete, FAIS Representative and Training Specialist at Bidvest Life.