Start

Two banks, one message, and what it means for the rest of us

In the space of 24 hours this week, the chief executives of HSBC and JPMorgan Chase, the largest bank in Europe and the largest in the United States, both stood in front of their own people and said something striking. Not striking because it was new. Striking because they said it out loud.

AI is going to take some banking jobs. Get ready.

Georges Elhedery, addressing HSBC’s 200,000 employees at an investor day in Hong Kong on Wednesday, put it plainly: “We all know generative AI will destroy certain jobs and will create new jobs.” His priority, he said, was making sure staff had “the capabilities, the training, the tools to make themselves future ready”. He cautioned against becoming “disenfranchised, anxious, overwhelmed, and resisting the change”.

Twenty-four hours later in Shanghai, Jamie Dimon was more direct still. JPMorgan, he told Bloomberg, would “be hiring more AI people and fewer bankers in certain categories”. He added, simply: “I think it will reduce our jobs down the road.”

You can read those statements as alarming, or you can read them as honest. I read them as both.

What strikes me is not really the headline. We’ve all heard versions of this since 2023. What strikes me is who’s saying it, and to whom. The chief executives of two banks with the deepest pockets, the largest training budgets, and the biggest in-house AI teams in the industry are quietly telling their own people that they need to take responsibility for their own development. The message has shifted from “we will look after you” to “you must adapt”.

If HSBC is saying that to staff with access to internal training programmes, internal AI tools, a brand-new chief AI officer appointed in March, and a global mobility programme, what should the self-employed mortgage broker in Macclesfield, or the case manager at a lender in the Midlands, take from it?

The answer, I think, is uncomfortable but useful.

Nobody is coming to retrain you.

There is no Elhedery in our world standing on a stage telling brokers what AI tools to use and which parts of the job will change. The trade press writes about lenders’ AI strategies. The networks run the occasional webinar. But the responsibility for staying relevant in a mortgage market that is already being reshaped, sourcing systems with AI prompts, criteria search tools that read free-text, distributors using AI to draft case notes, brokers piloting AI suitability checks, sits squarely with the individual.

This is where I want to be careful. There’s a tendency in pieces like this to either catastrophise or hand-wave. Neither serves anyone. The mortgage market is not about to be automated end-to-end. The conversation with a complex first-time buyer, the genuinely difficult later-life lending case, the self-employed application with three years of erratic accounts, these are not problems a large language model can solve on its own, and probably won’t be for a long time. The advice itself, the bit Consumer Duty most cares about, is still deeply human work.

The work around the advice, though, is changing fast. Affordability research, criteria checking, case preparation, KYC, document collection, follow-ups, even the first draft of a suitability letter, these are all being absorbed into AI tools that did not exist in usable form 18 months ago. If you spend most of your week on those tasks rather than on advising clients, the question is not whether your role will change. It is when.

Matthew Syed wrote a book a few years ago called Black Box Thinking. His argument, roughly, is that the difference between industries that improve and industries that stagnate is whether they look honestly at their failures and learn from them. The opposite of black box thinking is the closed-loop mindset: the refusal to acknowledge that anything needs to change because acknowledging it would be uncomfortable.

There is a parallel in how individual professionals respond to AI. The closed-loop response goes like this. “The regulator won’t allow it.” “Clients want a human.” “It’s all hype.” “I’ve heard this before with robo-advice.” Each statement contains a kernel of truth, which is what makes them dangerous. They give us permission not to look honestly at what is changing.

The open-loop response is less comforting but more useful. It accepts that some of what we do today will be done by software within five years. It then asks what, specifically. And then the harder question. What am I going to do this month to make sure I am still adding value when that happens?

Notice the time frame. Not “what is my five-year AI strategy”. That is a question for the network and the lender. The individual question is much smaller and much more uncomfortable. Have you used a single AI tool in your workflow this week? Did you put 30 minutes aside to learn one new thing about how generative AI is being deployed in lending? Did you ask a colleague what they’re trying?

The reason that small unit of time matters is that meaningful skill shifts almost always happen this way. Modest daily practice that compounds. The brokers who thrive in 2030 will not be the ones who shouted loudest about AI in 2026. They will be the ones who quietly started using it, badly at first, then more confidently, until it became second nature.

There is one more thing worth saying, because the banking CEOs didn’t, and they should have. AI literacy is also a defence. If you can’t sense-check what a tool produces, you’re at the mercy of its mistakes. In a regulated industry that puts the responsibility for advice firmly on the human, that is not a small problem. The brokers who understand how these tools work will be in a far stronger position to spot when they go wrong than those who treat them as a black box.

Elhedery told HSBC’s staff not to become “disenfranchised, anxious, overwhelmed”. Good advice. Switching off and hoping won’t dispel that anxiety. Engaging with the technology directly, on your terms, before someone else’s terms become the only ones available, just might.

If the people running the largest banks in the world are publicly telling their own staff to start adapting now, with all the support those institutions can offer, what does it say about the rest of us still waiting to see what happens?

Sources