AI’s impact on operations has been brutally efficient
Operations is the part of the financial services industry that makes things happen. That can mean a few things depending on whether they’re servicing an M&A or a sales & trading team, but generally speaking they are the machine that turns agreements into realities. The operations team, therefore, fulfills an exceptionally procedural role in the financial services process. Procedural roles are the exact sort that can be automated by AI – and are being automated by AI.
Click here to join the bubble by eFinancialCareers, our new anonymous community. ✍️
“The operations team is at the tip of the spear on using and leveraging new AI tools and capabilities,” said Marianne Lake, then-JPMorgan CEO of consumer and community banking (CCB), at the bank’s latest investor day, back in May last year. “Based upon what we know today, we expect headcount will trend down by about 10% over the next five years or so,” Lake said.
Lake even described that as a conservative number. "I would take the over on this projection," she said. “I'll bet we will deliver even more as the tools and capabilities just keep getting better and better.”
It’s not just the CCB. The firm’s CFO, Jeremy Barnum, also noted that JPMorgan was telling managers to “resist headcount growth where possible,” and to increase “their focus on efficiency.” That means investing in and using AI tools.
That was over a year ago, and AI has become much more capable since then, especially as an agent. In April 2026, the New York Times reported that JPMorgan, Citi, Bank of America, Goldman Sachs, Morgan Stanley and Wells Fargo had collectively shed around 15,000 employees while booking $47bn in profit in Q1 of 2026, up 18% on Q1 of 2025. All six credited AI to some degree, in areas running from the back office to the front.
And at a Goldman Sachs conference in December 2025, Wells Fargo CEO Charlie Scharf called AI "extremely significant", both for efficiency and for headcount. He even noted that most of his peers were reluctant to say so out loud.
The primacy of AI will be difficult to reverse, even if banks would want to (and they don’t). The UK, EU and Swiss markets move to T+1 settlement on 11 October 2027. This means that a securities trade must settle one business day after it is executed, rather than the current two. That means half the time to fix an error.
In the short term, that could be good for operations hiring, as the UK’s FCA noted in a blog post earlier this month that implementation was all over the place, but in the long term, it means that new problems must be solved quickly, and speed is a natural talent of an AI agent.
How does this look in practice? BNY, the custody firm, might offer an answer. BNY was a keen and early adopter of AI technology via its relationship with OpenAI. The result of that partnership was Eliza, a chatbot trained on BNY’s data.
Aside from acting as an ad hoc consultant and building apps for internal use, OpenAI itself noted that BNY had developed “digital employees” with names, access controls, and dedicated roles. It listed, as an example of “digital employee” work, the task of payment instruction validation, which is a core operations role – essentially, verifying the names and destinations of a payment are are correct before proceeding with it.
In turn, BNY’s staff is training Eliza and her brood. “Instead of handling certain tasks in the first instance, the role of the human operator is to be the trainer or the nurturer of the digital employee,” said BNY’s Chief Data and AI Officer, Sarthak Pattanaik. Operations professionals might not be out of a job immediately, but like in compliance for example, they are rapidly becoming supervisors, even if they weren’t senior before.
One area of operations that is particularly impacted by the potential of AI is clearing and settlements. These refer to the actual process, after a transaction, of transferring the financial product that was bought or sold to its new owner. Clearing and settlements are being transformed by tokenisation, which in turn makes it easier to implement AI. Tokenisation is, at its core, an ultra-efficient way of storing data, in which every asset that a bank deals with is turned into a “token” on a digital ledger. Instead of banks waiting days for paperwork to clear between different institutions, everyone in a transaction looks at the same digital token on the same ledger. This ledger, being digital and standardized, is then extremely easy to read and manipulate by an AI agent.
Some of the regions hit hardest will be traditional near/offshoring regions such as Poland, India, and Texas. These countries were chosen because they were cheap, had well-educated professionals, and a sufficiently trustworthy corporate culture to be trusted with standardized and automated processes such as onboarding clients.
It’s no surprise, therefore, that Citi’s cuts this year landed hard on cities like San Antonio, Tucson, and Tampa. Banks seek to cut costs, and Tampans are cheaper than New Yorkers. Unfortunately, Claude is cheaper than both.
Have a confidential story, tip, or comment you’d like to share? Contact: +44 7537 182250 (SMS, WhatsApp or voicemail). Telegram: @SarahButcher. Signal: sarahbutcher.22 Click here to fill in our anonymous form, or email editortips@efinancialcareers.com.
Bear with us if you leave a comment at the bottom of this article: comments are moderated intermittently by human beings. Sometimes these humans might be asleep, or away from their desks, so it may take a while for your comment to appear. You must take sole responsibility for comments you post on this site. We will take reasonable steps to weed out anything that we consider to be offensive or inappropriate.
