After a layoff · 5 min read

My job was cut because of AI. What can I do next?

If automation changed your job, map the tasks that disappeared, the work that remains and the roles where your experience still matters.

Written by Nivaan, founder of OffLadder · Last reviewed 10 October 2026 · Sources listed at the foot of this page

A job title is a bundle of tasks

Take a support lead. A model might draft replies and summarise tickets. It may be much less dependable when a vulnerable customer is distressed, a policy conflicts with another policy, or a failed process needs to be redesigned. The company can still choose to reduce headcount. But the employee's experience does not disappear with the position.

The ILO's 2025 analysis found that roughly one in four jobs worldwide has some exposure to generative AI. It emphasised that transformation is more likely than complete automation in many occupations. Exposure is not the same as a confirmed redundancy rate.

Make a map of the work that changed

List 12 tasks from your actual week. Put each in one of four groups: already automated reasonably well; accelerated by AI but still reviewed; depends on human relationships or accountability; and unclear or changing. Then circle the tasks where you were especially good. The overlap between useful human judgement and actual customer needs is more valuable than a slogan about 'human skills'.

Look for settings where that judgement travels. A recruiter might evaluate AI-assisted hiring processes. A writer might build editorial quality tests. An operations manager might redesign handoffs between software and frontline teams. These are directions to investigate, not guaranteed vacancies with standardised titles.

Separate the employer's explanation from the actual change 'AI' may be the label attached to a restructure, a cost-cutting decision, a tool rollout or a genuine collapse in demand for a task. Those mechanisms have different implications. Ask, where possible, which workflows changed, which outputs are still being delivered, who now checks errors and whether remaining colleagues are doing different work. You may never get a clean answer. But avoid turning the announcement into a verdict on everything you know.

A useful map has five columns: task; current automation; failure risk; human judgement; next employer who needs it. Take an experienced editor: generating alternate headlines has become easier. Deciding whether the headline misrepresents a sensitive story remains a consequential judgement. That judgement might be valuable in regulated communications, content assurance or AI evaluation, but only if real employers need it and you can show the work.

Look for the uncomfortable part of an AI workflow

The interesting opportunities are often where the demo stops: a customer gives ambiguous information, the source data is wrong, two policies conflict, the AI makes a confident mistake, or the team cannot agree who owns the decision. These are not magically automation-proof areas. They are places where experience may travel into the new way of working.

To investigate, read three job descriptions for relevant roles and underline verbs: review, escalate, investigate, map, train, assess, coordinate. Then build a sample. For a customer-support background, draw an escalation policy for a chatbot handling billing disputes. For legal support, create a checklist for checking the citations in generated case summaries. For marketing, design a rubric that scores factual accuracy, voice and substantiation. Never upload private employer or customer material to a public tool.

What if the new direction also changes? It will. 'AI workflow designer' may not be a standard job title in your city. Search for the underlying work across operations, product, quality assurance, process improvement and enablement roles. Check the local market. A credible move might be joining an ordinary operations team that is introducing AI rather than chasing a futuristic title with no vacancies.

Your next move should improve two kinds of resilience: the ability to work with changing tools and the ability to spot the limits of those tools. A course in prompting may help; a demonstrated improvement to a real process is stronger evidence. If you discover that you dislike supervising automated systems, that matters too.

Where OffLadder fits

OffLadder does not claim to know which jobs survive AI. It helps you investigate possible directions, including work still taking shape, through short experiments. Choose the part of your former work that you still find interesting, explore adjacent directions and try one small task. Laddie can help you examine a live route in and refine the next test; You keeps what you actually learned. Begin at https://offladder.com/. Use labour-market sources and local qualifications to check any promising title before committing money.