Generative AI is already reshaping white-collar work, but the deeper disruption in finance may not be the job losses everyone fears — it may be the bonus erosion nobody sees coming. Academic research shows AI’s productivity gains flow disproportionately to less experienced workers, compressing the performance gap between stars and journeymen. Banks including Goldman Sachs and Morgan Stanley are now pushing to capture and systematise the collective intelligence of their entire workforces. Bloomberg Opinion columnist Paul J. Davies examines what happens to pay, promotion and purpose when the machine knows everything your best banker knows — and never sleeps.
By Paul J. Davies
Morgan Stanley was among the first major banks to roll out generative artificial intelligence for its people to use in their day-to-day dealings with customers in late 2023. The internal chatbot was designed to give its financial advisers swift access to all the firm’s research and knowledge so they’d have answers or ideas on investments, tax or most other questions from wealthy clients. The point was to make every adviser as good as the smartest.
Now, the bank and several of its peers are pushing towards the next stage — how to capture and exploit all the information known and gathered by bankers as they talk with clients, strike deals or make trades. If successful, I suspect this will have profound effects on finance jobs. Even sophisticated roles dealing directly with clients could become much less intellectually creative than they seem today — and their oversized pay packages will likely shrink, too. In finance, AI might not steal your job but it could cut your bonus.
The coming change also poses fascinating questions about how high performers and future leaders can stand out – and whether investment banks will need to keep refreshing their workforces with traditional yearly rounds of firing and hiring.
Big investment banks have been trying to farm collective intelligence in human-led ways for years already; every strategy update involves a pledge to get different divisions to share information about clients so the bank can sell as many products as possible to each of them.
David Solomon launched his “One Goldman Sachs” push when he became chief executive officer in 2018 for just this purpose. Last year, as the bank’s leadership realized how AI could help redesign everything it does, it relaunched the effort as “One Goldman Sachs 3.0.” Morgan Stanley similarly has a vision for integrated client intelligence. Other major banks also have plans.
A key aim for all these efforts is to make all bankers as good as the best ones, just like Morgan Stanley’s much earlier wealth management tool was designed to do. The principle does seem to work, according to academic studies. Where AI has augmented jobs and helped people with tasks, in all kinds of white-collar work, not just finance, the productivity gains are concentrated in those with the least experience, according to Golo Henseke, an associate professor in applied economics at University College London. Generative AI seems to reduce the variation of productivity or output among people within the same occupation. In the past, new technologies often meant learning new skills, which some workers were better at doing than others.
The implications are quite different from what most people talk about when discussing the effects of AI on jobs in finance and elsewhere. Most attention has been on its power to take over mundane and repetitive tasks — reformatting pitchbooks, producing a first run of research or reports, doing administrative tasks, and so on. This would be systematic farming of the knowledge within a firm by the firm.
To understand what this might be like, think about a typical sales or advisory job. You keep up to speed with in-house research and activity, speak to clients about what they’re doing or might want to do, and try to combine this continually into ideas that you can sell to specific customers. The best salespeople produce more pitches that are better suited to their targets than others do. But if all that research and client intel is captured by the firm’s AI every moment of every working day, then the software will synthesize that knowledge into pertinent ideas faster than any salesperson. More of the job could become following instructions from your AI on who to call and what to pitch.
Bosses love this idea: Their staff will spend more time selling products that are definitely relevant and so more likely to bring in fees. But for the bankers, losing the creativity of coming up with ideas will make work more drone-like. And if everyone is fueled by the same intellectual resource, how do smarter people stand out, get paid more, or promoted and eventually emerge as the next generation of leaders?
“Maybe the game shifts to less AI-augmentable skills, like politics, networking and relationships, and human management,” says Henseke.
In finance, the higher up you already are and the more you control the relationship with a corporate client or fund manager, the more likely your role as it is will be safe. Other bankers who might stay ahead of the coming AI tsunami will be in jobs that involve more complex combinations of specific knowledge or skill. This is an extension of another idea bankers are already familiar with: You make more money from arranging customized derivatives trades than from buying and selling ordinary stocks or bonds, for example.
David Autor, economics professor at Massachusetts Institute of Technology, has studied how new technology effects occupations by assessing the expertise they still require after routine tasks have been automated. The research is based on how much technical jargon is needed to describe a job. For example, “cleaning windows” is simple language; “recalibrating the magnetic resonance imaging machine” is not.
The good news is that, historically, more expertise has meant greater pay after automation for most jobs since 1980. The one entertaining exception is for managers and executives where compensation seems to decline with this measure of expertise. That either means the less you know the more you earn, or that those bosses who are able to ditch jargon and speak most plainly are better paid, which would be a good answer.
If AI is going to equalize and improve the skillsets of most bankers, the final question is about how often people get fired for underperforming and make room for the next crop of bright young analysts. Traditionally, investment banks would cut the weakest 10% or so of revenue-earning staff each year and hire fresh graduate trainees. If AI makes variations in productivity disappear, there will be fewer weak performers to cull — and less need to bring in new blood.
Perhaps part of the reason for the slowdown in hiring of young people in many industries could be that AI is already improving the output of those already in the jobs. Researchers like Henseke say the data in the UK, for example, show a sharper drop off in hiring for occupations most likely to be altered by AI, but it’s too early to be sure of the reasons. “The pattern is there but we still have to be cautious about exactly what’s driving that,” he tells me.
Read more:
For bankers – and almost everyone else – the main AI survival strategy is to become as good as possible at using the new tools you have and try to ensure that the rest of your job is hard for anyone else to do. Or, to accept that your workday is going to become less creative — and probably pay a much smaller bonus.
© 2026 Bloomberg L.P.
Sign up for your early morning brew of the BizNews Insider to keep you up to speed with the content that matters. The newsletter will land in your inbox every morning on weekdays. Register here.
Support South Africa’s bastion of independent journalism, offering balanced insights on investments, business, and the political economy, by joining BizNews Premium. Register here.
If you prefer WhatsApp for updates, sign up to the BizNews channel here.
Leave a Reply