AI and work · 6 min read
What are AI agents, and will they take jobs?
What AI agents are, how they differ from chatbots, which tasks they're taking on, why many agent projects fail, and what it means for your job.
The short answer
AI agents are AI systems that can take actions to complete multi-step tasks, such as searching, filling forms, updating systems or writing and running code, rather than just answering questions. They're starting to handle structured, repeatable workflows in customer service, IT support, sales operations, coding and back-office processes. But they make mistakes, need oversight and are expensive to build well: Gartner predicted in 2025 that over 40% of agentic AI projects will be cancelled by the end of 2027 due to cost, unclear value or weak risk controls. For most workers, agents will change tasks before they change jobs.
Key takeaways
- Agents take actions across systems, not just answer questions.
- Structured, repeatable workflows are most affected.
- Many agent projects fail because of cost, reliability and risk.
- Supervising, designing and checking agents are growing skills.
On this pageWhat agents can do now
What agents can do now
| Area | Example tasks | Human role |
|---|---|---|
| Customer service | Resolve simple requests, process refunds | Handle exceptions and complaints |
| IT support | Reset passwords, troubleshoot common issues | Complex problems, security |
| Software | Write, test and fix code in steps | Review, design, accountability |
| Sales operations | Research leads, update CRM, draft emails | Relationships, negotiation |
| Back office | Process invoices, check documents | Exceptions, controls |
What agents struggle with
- Messy, ambiguous tasks with unclear goals.
- Situations not seen before.
- Tasks where mistakes are costly and hard to undo.
- Work requiring trust, empathy or physical presence.
How work will change
For most jobs, agents will take over parts of workflows: the repetitive, structured steps. People will spend more time handling exceptions, supervising agents, improving processes and doing relationship and judgement work. Some roles made mostly of structured digital tasks will shrink. New roles will appear around designing, testing and governing agents.
New roles around agents
- AI operations specialist: monitoring and improving agents.
- Workflow designer: mapping processes for automation.
- AI quality and evaluation specialist.
- AI governance and risk specialist.
Beware of agent washing
Gartner warns of 'agent washing': vendors relabelling existing chatbots and automation as agents. When your employer adopts an agent, ask what it actually does on its own, how errors are caught and who is accountable.
Illustrative example: an agent in customer service
An online retailer introduces an AI agent that can look up orders, process simple refunds and rebook deliveries. Within three months it handles about half of incoming chats end to end. The customer service team shrinks through natural turnover rather than layoffs, and the remaining agents handle complaints, complex cases and vulnerable customers. Two team members move into new roles reviewing the agent's conversations and improving its instructions, and one becomes the team's AI quality lead.
The lesson is that the jobs most at risk are the ones made entirely of the agent's tasks, and the people best placed are those who understand both the work and the agent well enough to improve it.
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:
- Explore: directions your experience carries into, including work that's only now taking shape as AI changes jobs.
- Try: an experiment small enough to finish this week, a small piece of the work itself.
- Learn: note which part you liked and which part you avoided, and keep it as a record of what you tried.
- 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's the difference between a chatbot and an AI agent?
- A chatbot answers questions in conversation. An agent pursues a goal by taking actions, such as looking up information, using software and updating records, often over several steps with limited human input.
- Which jobs are most exposed to AI agents?
- Roles built on structured digital workflows: customer service, IT helpdesk, data entry, scheduling, some sales operations and back-office processing. Roles needing physical work, trust, complex judgement or relationships are less exposed.
- Why do agent projects fail?
- Common reasons include high costs, unreliable performance on messy real-world tasks, unclear business value, security risks and lack of oversight. Many tools are also marketed as agents without real agent capabilities.
- How can I prepare for AI agents at work?
- Learn how agents work, volunteer to test them in your team, document your workflows clearly, and build skills in checking, handling exceptions and improving processes.
- Will AI agents cause mass unemployment?
- Forecasts differ widely. Most research, including the WEF Future of Jobs Report 2025, expects significant job change, with some roles shrinking and others growing. The pace depends on cost, reliability, regulation and how organisations adopt agents, which so far has been slower than vendor claims.
- Should I learn to build AI agents?
- If you're in a technical or operations role, learning the basics is useful: how agents use tools, how they fail, and how to evaluate them. For most people, understanding how to work alongside agents and spot their mistakes is more important than building them.
- How are companies using AI agents today?
- Mostly in contained areas with clear rules and easy checks: customer support for simple requests, IT tickets, coding tasks, document processing and research. Many keep a human approval step before any action that's costly or hard to reverse.
Exercises are practice, not evidence of career fit or a qualification. How to read our guides and report a correction.
Sources
- Gartner predicts over 40% of agentic AI projects will be canceled by end of 2027 · Gartner
- Future of Jobs Report 2025 · World Economic Forum
- Generative AI and Jobs: a global analysis of potential effects on job quantity and quality · International Labour Organization
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.