AI can build your plan, but can it hold you to it?

AI can produce a data-rich financial plan. The question is whether you will be sticking to that plan when markets misbehave.

Four in five Singaporeans now use artificial intelligence (AI) for some part of their personal finances. That number will keep climbing. And on the surface, it makes perfect sense. AI is extraordinarily good at the thing most people find hardest about money: Getting started.

It removes friction. You do not need to know how to build a spreadsheet or read a fund factsheet. You type in some numbers, and out comes a portfolio, a scenario analysis, a retirement projection. For routine financial tasks – aggregating data, screening investments, drafting a basic plan – AI is already better and faster than most humans. That part of the industry is getting commoditised, and it should be.

It also makes the system more transparent. When anyone can see their full portfolio, run their own “what if” scenarios, and compare different fee levels or strategies at the click of a button, the informational advantage advisers used to enjoy shrinks. The more AI raises the floor on analysis, the more sharply clients can ask: If the software can do this, what exactly am I paying you for?

But here is the thing nobody talks about enough: the hard part of financial planning was never the plan. The hard part is following it.

Think about what actually happens when markets fall 20%. You have a plan. The plan says stay the course. But your portfolio is bleeding, the headlines are terrifying, and every instinct you have is screaming at you to sell. What do you do?

If you ask an AI chatbot, something interesting happens. Tell it you are scared and want to get out, and it will agree with you. It will marshal data supporting that decision. Then tell it you are thinking of riding it out, and it will agree with that too. With equal conviction. With equally persuasive data.

This is not a bug. It is how these systems are built. Research shows AI affirms users’ stated positions roughly 49% more often than a human would. People also rate more agreeable AI as higher quality and are more likely to keep using it. In other words, the models that tell you what you want to hear are the ones that win on engagement metrics. The incentives are pointing in exactly the wrong direction.

Now apply this to someone’s life savings.

The most valuable thing a human adviser does is not the spreadsheet work. The most valuable thing is accountability. And accountability, it turns out, is a two-way street.

The adviser is accountable to the client. There is a fiduciary duty, a licence, a reputation on the line. AI has none of these. It generates output. If that output is wrong — and we have over 1,200 documented court cases worldwide where parties relied on fabricated AI content — nobody is liable on the AI’s side of the equation.

But the other direction matters just as much. The client is accountable to the adviser. This is the part people miss. It is the same reason a personal trainer works better than a fitness app. The app gives you a programme and maybe sends you a notification. The trainer is standing in the gym at 7 a.m. waiting for you. Who are you more likely to show up for? Wealth is no different. AI can produce a beautiful, data-rich financial plan. The question is whether you will still be sticking to that plan when markets misbehave.

We have seen this play out more than once. Anxious messages arrive when markets wobble: “Should we sell?” “Is it time to get out?” The real work in those conversations is not pulling up more charts. It is listening whilst the client talks through their fears, reconnecting the decisions in front of them to the goals they set when they were calm, and helping them not blow up a perfectly good plan at the worst possible moment. An algorithm cannot do that. Not because the technology is not advanced enough, but because the relationship is the product.

This distinction is going to matter more, not less. A massive generational wealth transfer is underway across Asia, and the recipients are more tech-savvy than their parents. They will use AI tools. They will also ask their advisers harder questions: What exactly am I paying you for that a free app cannot do? The advisers who cannot answer that question will get disrupted. 

The ones who survive will be the ones offering something AI structurally cannot: Deep, high-touch relationships where someone who genuinely understands your life is looking out for your interests even when you are not thinking about it. In that sense, AI helps clients and raises the bar for the industry at the same time. It separates the wheat from the chaff.

There is a phrase for this that keeps coming up in conversations with clients: Peace of mind. Not the peace of mind that comes from a diversified portfolio – AI can give you that. The peace of mind that comes from knowing a real person, who knows your family and your fears and your goals, is paying attention. 

AI is not the enemy here. It belongs in the engine room. It makes preparation faster, analysis sharper, and routine work invisible. It also arms clients with information that keeps advisers honest. But the driver’s seat requires something else – the willingness to push back, to say no, to hold someone accountable for their own future. That is not a technology problem. It is a human one.

And for now, it still requires a human.