What impact is AI having on sales & trading careers?
When people talking about sales & trading careers, and the skills required for them, they generally tend to take a classical interpretation - phone calls and steak dinners shouting, oversized phones. Pocket squares. But AI is changing everything, and sales & trading is no different.
Click here to join the bubble by eFinancialCareers, our new anonymous community. ✍️
Back in December, we hosted an AMA on our forum for financial services professionals - the bubble - with a sales trader with over 20 years of experience. "Half the roles done by humans don't need humans, not any more," he said. The remaining tasks in sales trading include: "admin, report compiling, data pulls, collation of news [and] writing summaries,” he said.
These are mostly middle- and back-office functions which support the salespeople and traders, but they are what junior traders are often trained upon. A London-based salesperson we interviewed in June 2026 noted that her job involved a lot of these administrative tasks.
Unfortunately, these are the kinds of things that are now susceptible to AI. “Over the past decade, a big chunk of generalist equity sales roles have disappeared," the sales trader on the bubble said.
A BCG report from April this year noted that 70% to 80% of manual workflows for traders could be automated – and, practically, BCG noted that this means “reducing friction across request-for-quote handling, negotiation support, mandate interpretation, documentation review, compliance checks, and transaction workflow management.”
Financial services firms can profit from the automation and efficiency that AI offers. Norges Bank, the Norwegian central bank, noted that the firm had reduced trading costs by 30%, or approximately $500m, between 2023 and 2025. It credited AI adoption with the reductions.
Put simply, anything that doesn't involve speaking to a client can be automated to varying degrees. Automation means electronification, or trading electronically. And electronification opens the door to the use of AI.
Electronification is most often used in standardised markets with high volumes of trades (high liquidity), like FX and basic equities. However, AI can plug the gaps in markets where products are less standardised, like high yield bonds where the underlying resilience of the company issuing the bonds is unclear. BCG estimated that AI’s role in “less electronic” markets could impact 30 to 40% of the volume that a typical multi-asset desk deals with. In high yield for example, Jim Kwiatkowski, CEO of LTX, a company digitizing bond trading, told The Trade that AI could increase market liquidity by expanding the "universe" of bonds that traders can "analyze and consider when making decisions," and by revealing new potential counterparties for those bonds.
Sanjay Jhamna, JPMorgan's global head of credit trading, told Bloomberg earlier this year that AI will "reset" who can compete in credit trading because it can create prices out of patchy or unstructured data, a traditional weakness of electronification models.
Generally, the fields safest from AI are the ones safest from electronification. For example even though AI can be used to help price high-yield credit, trades are often tailored to a particular client's risk appetite. Personal relationships can smooth over those cracks. Even in algorithmic trading, though, one senior trader tells us humans will always be needed to monitor what AI is doing.
It’s therefore unlikely that AI will ever fully replace sales & trading roles. As the equity sales trader in the bubble said, “clients will always make time for someone who knows their coverage deeply, can challenge their thinking, and can spot dislocations or opportunities early." AI cannot replace relationships, and it won’t ever be able to (well, except romantic ones).
Banking is a heavily regulated industry. The European salesperson that we spoke to told us that while her bank automates many of its processes, and was constantly endeavoring to automate them, she had minimal contact with agentic AI tools. Clients that have complicated requests expect a human to answer emails, calls, and Bloomberg requests.
Perhaps more worryingly for buy-side traders, who are the clients that banks buy and sell securities on behalf of, AI models have proven themselves to be quite savvy investors, too. JPMorgan staged an important experiment earlier this year, Bloomberg reported, in which the firm set up AI agents to trade hypothetically based upon real data from the past. In many cases, the AI models beat traditional 60:40 investing strategies, which are the usual ones offered to private investors: a portfolio comprising 60% equities and 40% bonds.
Using past trading data to look at how a trading hypothesis might perform is called backtesting and is critical to trading using computer algorithms. Backtesting can be done using AI.
JPMorgan backtested its AI models on two decades of market performances. The implication of its study is that novel techniques and investment strategies can be tested this way, too. And AI can run thousands of tests, far more quickly than human beings.
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.
