Like many others I have watched the rapid rise of artificial intelligence (AI) with a mixture of fascination and unease since we were first introduced to ChatGPT. What began as curiosity quickly became more personal: what does this mean for my role? What does it mean for my business?
For a while I stayed out of the debate. Not because it was irrelevant, but because I was processing it. AI felt powerful, but slightly distant. It had not yet meaningfully intersected with my day-to-day work as a financial adviser.
That changed recently. I was sitting across from a prospective client when the implications of this new world became very real.
“I’ve already asked AI what I should do,” she said.
She had run the scenarios herself. Worked through the asset allocation and tax implications. Modelled her retirement projections. The system had produced a clear recommendation: rebalance the portfolio, increase equity exposure, realise some gains and optimise the overall structure.
On paper, it was entirely sensible. And yet she hesitated. On paper, the recommendation made sense. It was rational, structured and directionally sound. But that was not what was holding her back. She was uncomfortable. That hesitation tells you everything you need to know about the debate around AI and financial planning.
It is easy to model the tax consequences of moving money offshore. It is far harder to unpack the emotional weight of deciding whether you believe in the country in which you built your life
If this profession were purely about calculation, it would already be obsolete. AI can explain retirement annuity limits in seconds. It can calculate capital gains tax instantly. It can model a 30-year retirement projection with speed and precision that few humans can match. It does not forget legislation; it does not get tired and does not rely on memory.
For optimisation, it is exceptional. But financial planning is not fundamentally an optimisation problem. It is a behavioural one. The industry keeps asking whether AI will replace advisers, but that is the wrong debate. The more important question is not who wins between human and machine, but what problem financial planning is trying to solve in the first place.
If the problem is access to information, AI wins comfortably. Clients have tools that can generate technically sound answers in seconds. If an adviser’s value lies only in explaining tax rules or constructing a generic balanced portfolio, that value will be commoditised. But most investors do not derail their plans because they misunderstood a tax rate. They derail them because they reacted emotionally at the wrong time.
They panic after markets fall. They become overconfident after markets rise. They resist realising gains because paying capital gains tax feels like a loss, even though the gain exists only because the investment worked. In a country where marginal tax rates reach 45%, that emotional resistance is amplified. The cheque to the taxman feels personal, even when the portfolio is materially better off.
In South Africa these decisions are rarely abstract. Investors are navigating political noise, currency volatility, debates about expropriation, load-shedding memories and constant discussion about whether to externalise assets. It is easy to model the tax consequences of moving money offshore. It is far harder to unpack the emotional weight of deciding whether you believe in the country in which you built your life.
These are not spreadsheet problems. They are human ones. Every client walks into a meeting with a money story. How they grew up. What money represented in their household. Whether it meant safety, control, status or survival. In a society as unequal and economically complex as ours, those stories are often deeply rooted. Experiences of scarcity, political transition, currency shocks and family pressure shape how people respond to risk and opportunity.
AI can tell you what is right. It cannot tell you what you are ready for
AI can tell you the statistically optimal equity allocation given your time horizon. It cannot sense that you lose sleep when the rand weakens sharply. AI can recommend rebalancing after a strong year on the JSE. It cannot feel the hesitation when selling a winning asset triggers capital gains tax and makes success feel like something is being taken away.
That gap between logic and emotion is where the real work of financial planning happens. The best advisers are not walking tax manuals. They are behavioural interpreters. They notice the pause when someone says they are comfortable with risk. They recognise tension between spouses when emigration is discussed. They understand selling a business is not only a liquidity event but an identity shift.
Often the technical answer is clear, but the emotional readiness is not. The best plan is not the mathematically perfect one. It is the one a client can live with. The one they will stick to when markets fall 20% and headlines become dramatic. A technically optimal portfolio abandoned during volatility is inferior to a slightly less aggressive portfolio that survives the cycle.
None of this diminishes the power of AI. In fact, it clarifies it. AI will make good advisers better. It will handle modelling, comparisons and legislative interpretation with speed and accuracy. It will remove inefficiency and will expose advisers whose value proposition was built on information asymmetry rather than judgment.
What remains is the human layer. The advisers of the future will not be defined by fund selection. They will be defined by judgement. AI will handle precision. The adviser will navigate the moments where emotion, identity and long-term conviction collide.
AI can tell you what is right. It cannot tell you what you are ready for. In a country where financial decisions are often intertwined with questions about security, opportunity and belonging, that difference matters more than ever.
• Marrian is director at independent wealth management firm InvestSense.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.