If you haven’t already, subscribe and join our community in receiving weekly AI insights, updates and interviews with industry experts straight to your feed.

DeepDive 

Your weekly immersion in AI 

If you walked into a major bank in 2012 and told someone your job title was ‘DevOps engineer,’ they wouldn’t have known what you were talking about. 

At the time, software teams, IT departments, and infrastructure managers all worked separately. Then cloud computing came along, and suddenly organisations needed people who could connect development, operations, automation, governance, and reliability into one coherent system.

A new discipline emerged almost overnight. And something similar might be happening right now with AI. 

A new category of work is emerging, crossing the boundaries between IT, compliance, operations, and strategy. The job titles in this space vary wildly – we’re seeing AI governance leads, model risk specialists, AI workflow architects, prompt specialists, AI transformation directors…but all of them are essentially tasked with making enterprise AI actually work, safely and at scale. 

For people starting their careers in tech, this means the next generation of AI jobs will require people who can coordinate complex human and machine systems simultaneously.

The rise of AI operations 

Everyone’s been talking about how AI could displace human jobs. Over the last two years, conversations have focused on which industries are most exposed, and which workers are most vulnerable. 

But inside organisations, business leaders are discovering that models don’t manage themselves. Someone has to…

• Evaluate outputs
• Monitor risk
• Maintain internal knowledge systems
• Review workflows
• Train employees
• Establish governance rules

…and ultimately decide when humans should stay in the loop.

So enterprise AI is creating operational complexity – and operational complexity creates jobs. 

At JPMorgan Chase, Reuters recently reported that parts of the bank were reorganised around data and AI strategy, including governance, infrastructure for AI agents, and workflow transformation initiatives.

Rather than a temporary innovation lab, that sounds like the early stages of a permanent enterprise function.

Meanwhile, Morgan Stanley has built internal AI systems for financial advisers that still rely heavily on human review and oversight. The firm’s AI assistant helps advisers retrieve and summarise information, but humans are still responsible for judgement, client communication, and final decisions. 

We keep seeing evidence like this that the future of AI work isn’t going to be a case of humans against machines – instead, it increasingly involves humans managing systems of machines. 

The jobs nobody talked about two years ago

We’re seeing demand for people who can: 

• Evaluate model outputs and hallucination risks
• Build governance and compliance frameworks
• Manage retrieval systems and enterprise knowledge bases
• Design human-in-the-loop workflows
• Train employees to use AI responsibly
• Coordinate AI systems across departments
• Translate between technical teams and business leaders

Some of these roles didn’t exist (or at least, not at scale) before generative AI went mainstream – and now they’re appearing everywhere. 

Many of them require hybrid skills. Instead of pure coding or pure strategy or pure legal expertise, organisations need combinations of a range of different capabilities. And that’s one of the reasons why regulated industries are moving so aggressively here – because organisations can’t just deploy AI and hope for the best. 

They need people to manage how those tools interact with real-world processes. 

At Moderna, the company expanded enterprise AI usage across business functions with hundreds of custom GPTs internally deployed to support teams across legal, research, and operations. 

And Klarna became famous for rapidly integrating AI into customer-service workflows and internal operations. But after scaling AI support systems, Klarna also found that some customer experiences still required greater human involvement and empathy.

This highlights a pressing reality: AI systems are (they have to be) managed environments. And as organisations deploy more tools, someone has to decide which tasks and decisions can be automated, and which ones need human review. 

These are operational questions. And increasingly, they’re becoming career paths. 

The next big enterprise function 

Cloud computing created DevOps because software became infrastructure. Generative AI may now be creating AI operations because intelligence itself is becoming infrastructure.

So to adapt quickly and successfully, organisations need to build strong operational systems around their AI models – including governance and human collaboration. 

If you’re wondering where you fit into the AI economy, this could open a wide door of career possibilities. Because the future of AI work needs translators and coordinators; auditors, strategists, and workflow designers; to help organisations use AI responsibly in the real world. 

What’s the most unexpected AI job ad you’ve seen? 

Open this newsletter on LinkedIn and tell us in the comments. 

We’ll see you back in your inbox next week.