The wealth manager nobody asked for: why an AI advisor beats a human with sales incentives

By Momentum IA · June 30, 2026.

There is a semantic distinction that Joyce Hooi, a columnist for Singapore's Business Times, turns into a central thesis: 'AI relationships' (emotional bonds with chatbots) leave her skeptical; 'AI relationship managers' (artificial wealth managers) strike her as a long-awaited liberation. The difference matters more than it seems.

The trigger for the article is an awkward call with the latest relationship manager—the term banks use for the advisors who serve their mass affluent clients—assigned to her account. The man tries to persuade her to meet in person. Hooi, with the clear-sightedness of someone who has already seen this movie, wonders what he will sell her this time: a structured note linked to equities? An investment fund with a front-end load? An investment-linked life insurance policy with high premiums? The pattern is familiar across retail private banking throughout Asia: the RM is paid, directly or indirectly, for placing product. Their interests and the client's do not align by default; at best, they align by chance.

The threshold Hooi mentions is revealing: if you have less than a million dollars in liquid assets, the bank is already considering handing you off to a chatbot. Far from being scandalized, the columnist writes that that day cannot come soon enough. It is a stance that many mid-tier private banking clients would share if asked honestly.

**The conflict of incentives that AI could break**

The underlying problem is not that human managers are incompetent or malicious—most are not—but that they operate within an incentive structure that pushes them to sell product rather than to advise. A language model managing a portfolio earns no trailer fee for placing a fund and has no monthly quota of structured products. Its role could, in theory, align with the only legitimate interest: that the client's wealth grows in a way suited to their risk profile.

As sector context, robo-advisors have been promising exactly this for over a decade, with partial success: they have democratized access to low-cost index portfolios, but have failed to provide the comprehensive financial planning component—inheritances, taxation, life changes—that a good human advisor can handle. The difference with the new generation of AI agents is one of orders of magnitude: an advanced LLM can hold complex conversations, process tax documentation, gauge the impact of a job change on a savings strategy and update the plan in real time. It is not a risk-profile questionnaire; it is something qualitatively different.

**Who wins and who loses**

Who wins is clear: mass affluent clients—the segment banks serve poorly because it is not profitable enough to devote quality human time to. Until now they received the worst of both worlds: human attention too oriented toward selling and too scarce to be genuinely personalized. An AI manager available 24 hours a day, with no sales quotas and full memory of the client's financial history, could be the first time this segment gains access to real advice.

The obvious losers are low- to mid-tier personal banking managers, whose profile—retail bank RMs with limited product access and basic financial training—has very little to offer against a well-trained AI agent. High-end wealth advisors—family offices, genuine private banking for large fortunes—have more room: they manage relationships, multijurisdictional tax complexity and situations where human judgment and personal trust remain hard to replicate. AI will displace them more slowly, if at all.

As sector context, there is a third actor to watch: regulators. Bodies such as the MAS (Monetary Authority of Singapore) are working on the use of AI in financial services, and the debate over liability when an autonomous agent makes a flawed investment decision is far from resolved. The legal framework for automated active-management advice—not just index portfolios, but dynamic strategies—remains a gray area in almost every jurisdiction.

**Two types of relationship that should not be confused**

Hooi is right to separate the two categories the headline sets against each other. Emotional relationships with AI—virtual companions, dating chatbots, synthetic friends—raise philosophical and mental-health questions that deserve their own cautious analysis: the risk of replacing real human bonds, emotional dependence, the manipulation of vulnerable users. It is not a territory where easy optimism is advisable.

Automated financial advice is something different: the provision of a technical service where human bias has historically been a problem, not a virtue. Here AI replaces neither affection nor community; it replaces a perverse incentive. They are morally distinct categories and should not be mixed.

In the long run, this transition is part of a broader reordering: AI as infrastructure that levels access to services once reserved for those who could pay for top-tier professionals. Legal, medical and financial advice—for decades distributed deeply unequally according to income. The horizon in which a client with modest savings receives wealth guidance of the same quality as a billionaire is not utopian; it is the direction the technology is pointing. The transition will be hard for those who today make a living off that asymmetry of access. But the net result—universal access to quality advice—is genuinely valuable.