← The Future of Work series The New AI Roles That Didn't Exist a Year Ago
Part of the Future of Work series — I've been tracking everything written about the future of work in the AI era. Research, decisions, predictions. A lot of noise. Some signal. In each post, I take one specific move and ask: what does this actually mean?

Job postings for roles that didn't exist two years ago are everywhere now. AI Ops Manager. Prompt Engineer. Human-AI Interaction Designer.

These aren't existing roles with "AI" slapped on. These are entirely new functions.

And nobody knows what they actually do.

The new roles emerging

Four new role categories we're seeing everywhere:

AI Operations Manager

This person manages the AI systems themselves. Not the people who build AI. Not the people who use AI. The AI.

They monitor LLM performance, manage API costs, optimize token usage, handle rate limiting, manage model versions.

It's like being a DevOps engineer, but for language models.

Prompt Engineer / Prompt Architect

Getting the right answer from AI isn't obvious. It requires knowing the model's quirks, building the right context, asking questions the right way.

Prompt engineers are the people who get consistent, production-grade outputs from AI systems.

It sounds weird. But companies are paying $80-120K for someone who can do this well.

Human-AI Interaction Specialist

How do humans and AI work together without friction? This role designs the interaction patterns.

When do you trust the AI output? When do you ask for human review? How do you escalate? How do you teach the human to work alongside the AI?

It's part UX designer, part behavioral psychologist.

Agentic Workflow Coordinator

This is the person who designs how AI agents and humans coordinate.

Who decides? The human or the agent? At what confidence threshold does the agent act alone? When does it ask?

It sounds abstract. It's actually very concrete: they're designing power dynamics.

If this plays out

Within two years, these four roles become standard in any organization running AI systems at scale.

Companies will fight over prompt engineers the way they fought over machine learning engineers.

Who wins

People who learn these skills now. People who can bridge technology and human behavior. People comfortable with ambiguity.

Who loses

People whose careers were built on "I know how to extract information from systems." That's now automated.

Want more on managing in the AI era?

The Little Book for New Managers has helped 30,000+ managers figure out what actually works.

Get the book ($19) →