AI relationships, no. AI relationship managers, yes

AS A happily married – if frequently exasperated – woman, I hadn’t expected to have the “It’s not you, it’s me” conversation with a man last week.

But there I was, trapped in a doomed phone call with the latest relationship manager to be saddled with my plodding mass-affluent bank account.

As he tried to cajole me into meeting in person, I wondered what he would attempt to sell me this time. An equity-linked note product? A unit trust with an upfront sales charge? An investment-linked insurance policy with hefty premium payments?

I made polite “I’m busy” noises until he finally hung up.

Maybe his overtures would be welcome and even beneficial to the next person on his call list. But people like me are an absolute nightmare for people like him, and there might be even more people like me in the future.

A recent HSBC survey found that while 62 per cent of wealthy people prefer human professionals as the main source of investment ideas, people in my age group – Gen Z and millennials – consistently preferred the hybrid approach, using both artificial intelligence and human advisers “across every major financial task”.

Mass-affluent clients

It doesn’t help either that my peers are likely in the mass-affluent bucket – defined as having liquid assets worth between US$100,000 and US$1 million. Most of the investment options available to us are generic, fungible and easily commoditised.

At the same time, my demographic is not guaranteed to join the ranks of the wealthy, but most of us will eventually enter older age brackets, bringing with us our aversion to answering phone calls from unknown numbers and our tendency to confide in large language models (LLM).

Eventually, the pool of mass-affluent customers willing to pick up their relationship manager’s calls will shrink. The banking sector must intuit this – consultancy McKinsey & Co recently implied that AI is likely to replace humans for clients with US$1 million or less in liquid assets.

“The mass-affluent client now gets something close to private-banking quality from AI,” a senior partner at McKinsey said.

But getting “private-banking quality” out of an LLM takes more than asking it for a stock tip. Where AI shines is in combing through vast amounts of verbiage and data to produce specific answers to targeted questions.

You could, for instance, run a structured product brochure through an LLM and have it look for hidden fees or generate worst-case scenarios for your principal.

You will need a basic understanding of personal finance to know the right questions to ask, but if you lack that, it’s possible to reverse-engineer your way to those questions by sourcing them across a range of chatbots and assembling the best ones.

There is still room for humans in banking, though. In discussing the results of its survey, HSBC reckons that what clients value in a wealth adviser is judgment, context and – most tellingly – accountability.

An LLM might eventually close the judgment and context gap with a human being, but it will not absorb accountability for your subpar portfolio.

For now, I’ll take my chances with the machines. A chatbot won’t provide a sympathetic ear when the market tanks, but at least it won’t disrupt my Tuesday afternoon to sell me an investment-linked policy.