According to Business Insider, technology leaders at major banks like Citi, Morgan Stanley, and Capital One are in a race to upskill their massive engineering teams for the AI era, with Citi and Morgan Stanley respectively employing 30,000 and 15,000 developers. Trevor Brosnan of Morgan Stanley says engineers are now in “continuous learning mode,” adapting to shifts that arrive every few months instead of every few years. A key part of this training involves developers learning to communicate with generative AI agents in English, not code—a skill that doesn’t come naturally to everyone in tech. Banks are deploying a mix of internal “AI Academies,” short video modules, and programs like Citi’s “Techflix” series to accelerate adoption. The goal is to use AI to boost efficiency and cut costs, but it requires a “fundamental shift” in a developer’s role from writing all the code to delegating larger tasks to AI agents.
The Fundamental Shift: English, Not Java
Here’s the thing that really stands out. For decades, the most valuable language in a developer’s toolkit was a programming language. Now, according to these execs, it’s becoming plain old English. Dov Katz at Morgan Stanley put it bluntly: “not everybody in technology has a reputation for being an excellent communicator.” And Citi’s CIO, Jonathan Lofthouse, joked that most developers’ second language is Java, which is sometimes easier for them to express problems in. That’s a huge cultural and skillset pivot. The job is morphing from being the sole coder to being a manager and spec-writer for AI agents. You’re not just asking for a snippet of code anymore; you’re delegating an entire task and describing the desired output with precision. It’s a whole different kind of problem-solving.
The Upskill Industrial Complex
So how are these banks, with tens of thousands of engineers, trying to pull this off? They’re building what looks like an entire internal education industry. Capital One has its “AI Academies,” which Nish Rana calls a “one-stop solution.” Morgan Stanley mixes outsourced and in-house courses. Citi ran “Techflix.” And all of them are leaning hard on short-form video content—some as brief as five minutes—recognizing that, like the rest of us, their employees are already learning from YouTube. They’re betting that these opt-in, flexible programs will work better than mandatory training. But let’s be real: this is also about defense. They need to protect their massive existing investments in human capital. Upskilling is cheaper and faster than firing 15,000 engineers and hoping to find 15,000 new ones with perfect AI prompt skills.
Anxiety And The Psychological Safety Net
You can’t talk about this breakneck pace of change without addressing the anxiety. The article mentions veteran engineers who are worried about keeping up, and the broader fear that a computer science degree is losing its luster when AI can spit out code. The banks are acutely aware of this. Rana at Capital One said a big goal of all this training is to provide a “psychological safety net.” It’s about creating a corporate sandbox where it’s okay to experiment and learn without the immediate pressure of production. That’s smart. Because if engineers are just terrified of being obsoleted, they’ll freeze up or leave. The promised land, according to the leaders, is that AI will free engineers from grunt work for “higher-order thinking.” But getting there is going to be a bumpy, stressful ride for a lot of people.
The Bigger Picture: What’s Really Going On?
Look, when banks pour billions into something, you pay attention. This isn’t just about cool tech. It’s a hard-nosed business calculation about productivity and cost. The dream is an engineer who, with a well-crafted English prompt, can get an AI agent to do the work that used to take a junior team a week. That’s how you get “efficiency” and “cut costs.” But there’s a fascinating tension here. They still insist foundational coding knowledge is critical because, and this is key, humans have to check all the code that AI writes. So the job isn’t disappearing. It’s changing into a hybrid role: part product manager, part QA analyst, part creative spec writer. The engineers who embrace this “continuous learning” mindset and get comfortable talking to machines like colleagues will probably thrive. The ones waiting for the next big shift to settle down might be waiting forever.
