Skip to content

New work in the AI era · 7 min read

Agentic AI jobs: who builds and runs AI agents

AI agents take actions, not just answer questions. The jobs that design, build, test and supervise them, the skills needed, and where to start.

The short answer

Agentic AI jobs are roles that design, build, test and supervise AI agents: systems that take actions such as sending emails, updating records, browsing, writing code or running multi-step workflows, rather than only answering questions. On the building side, AI engineers design agent workflows, connect tools and data, and write evaluations. On the running side, agent operators and AI operations specialists monitor what agents do, catch errors, handle exceptions and decide when to stop them. On the design and governance side, people decide which tasks agents should do, what permissions they get and how humans stay in control. Because agents act in the real world, mistakes cost more, so oversight, testing and permissions work is growing as fast as building.

Key takeaways

  • Agents take actions, so mistakes have real consequences.
  • Builders design workflows, tools and evaluations.
  • Operators monitor agents, handle exceptions and stop them when needed.
  • Permissions and human oversight are growing specialisms.
On this pageWhy agents create new jobs
  1. Why agents create new jobs
  2. The roles
  3. What good agent design looks like
  4. Operating agents is a real job
  5. Is it right for you?
  6. How to start
  7. Where the jobs are
  8. Illustrative example
  9. Test a future job before you train for it
  10. Questions people ask

Why agents create new jobs

A chatbot that gives a wrong answer is a problem. An agent that sends a wrong payment, deletes a file or emails a customer with false information is a bigger one. As organisations hand multi-step tasks to agents, they need people who design them carefully, test them thoroughly and watch them closely.

The roles

RoleWhat they doBackground
AI engineer (agents)Build workflows, connect tools, write evalsSoftware engineering
Agent operatorMonitor, review and correct agentsOperations, support, QA
Automation specialistConfigure agents in business toolsOperations, IT
AI security specialistTest injection and permission risksSecurity
AI governanceSet rules on what agents may doRisk, compliance
Agentic AI roles

What good agent design looks like

  • Start with a narrow, well-defined task.
  • Give the agent the fewest permissions it needs.
  • Ask a human before irreversible actions.
  • Log every action so it can be reviewed.
  • Test with tricky and adversarial cases before launch.

Operating agents is a real job

Organisations running agents at scale need people watching dashboards of agent activity, reviewing flagged actions, handling exceptions and improving instructions. It resembles supervising a team, with the difference that the team works very fast and doesn't get tired. OffLadder's direction agentic systems operator describes this work with a small experiment to try.

Is it right for you?

Building agents suits engineers who like systems thinking and edge cases. Operating them suits careful, calm people who notice when something is slightly off and enjoy improving processes. Both suit people who are comfortable with responsibility for things that move fast.

How to start

Build a tiny agent with a no-code platform or a framework: one task, two tools, a human approval step. Break it on purpose, and fix it. Write up what happened. When applying, OffLadder CV reads the post and shows which must-haves your CV doesn't show yet, such as a specific framework.

Where the jobs are

Customer service, IT support, software development, finance operations and sales are where agents are being deployed fastest, so that's where agent roles cluster. Search for agent engineer, AI agent developer, AI operations and automation titles.

Illustrative example

Nia, a customer support team lead, is asked to supervise a new AI agent that handles refunds. She spots that it approves refunds for duplicate accounts and builds a review rule. She becomes the company's first agent operations lead.

Test a future job before you train for it

A forecast can tell you which jobs are growing. It can't tell you whether you'd enjoy the work. Laddie, our AI coach, starts from the experience you already have and works through four steps with you:

  1. Explore: directions your experience carries into, including work that's only now taking shape as AI changes jobs.
  2. Try: an experiment small enough to finish this week, a small piece of the work itself.
  3. Learn: note which part you liked and which part you avoided, and keep it as a record of what you tried.
  4. Apply: when a direction turns into a job post, OffLadder CV shapes your CV for it. What you can't show yet goes back to Laddie as a short lesson, a 15-minute task and a mock interview.

Questions people ask

What is an AI agent?
Software that uses an AI model to decide and take steps towards a goal, using tools such as email, databases or browsers, with varying levels of human oversight.
What does an agent operator do?
Watches what agents are doing, reviews their actions, handles cases they can't, fixes failures and adjusts their instructions or permissions.
Do I need to code to work with agents?
Building custom agents usually needs coding. Configuring agents in no-code platforms and operating them often doesn't.
Will AI agents replace jobs?
They are taking on some multi-step tasks, especially in customer service, back-office work and coding. They also create work in building, supervising and governing them.
What are the biggest risks with agents?
Taking wrong or harmful actions, following hidden instructions in content they read, having too many permissions, and failing silently.
What skills are most in demand?
Designing reliable workflows, writing evaluations, setting permissions, monitoring and incident handling, and understanding the business process the agent works in.
Is this a passing trend?
The terminology may change, but software that takes actions on people's behalf is likely to keep growing, and so will the need to build and supervise it.

Share this guide

Exercises are practice, not evidence of career fit or a qualification. How to read our guides and report a correction.

Sources

About the author

Nivaan builds and writes OffLadder. His background is applied AI and enterprise work, which is where the AI and career-change guides come from. More about Nivaan.

Spotted something out of date or wrong? Email hello@offladder.com and we will check it and update the page.

How this page is written

Sources support the factual background, not a guarantee that an exercise predicts career fit. Fictional practice material is labelled. Dates record content updates; no independent expert review is implied. More about OffLadder.

The career explorer that gets you.

OffLadder is a career explorer that gets you and keeps up as you change. Laddie, its AI coach, tells you what it heard, suggests directions your experience carries into, gives you a piece of the work to try and learns from how it went, then finds open roles worth a look and shapes your CV for each job. Answer four quick questions, plus one optional, then open a free account with Google or just your email, no password, and Laddie writes your first read straight away.

Which is closest to you right now?

One tap answers that question for you. A few more and Laddie writes you a direction to test.

Find a direction to try