NexGenTekTalk to us
AI and automation

AI and jobs: which work changes, which stays with people, and how to prepare

A support team lead watches a new chatbot answer the password resets, order queries and account questions her team used to handle. Her team is smaller than it was last year, and she wonders whether her own role is next.

She is right that something real is happening, and wrong if she assumes the whole job will disappear. AI takes over tasks more often than whole jobs. Some roles will shrink, many will change, and the people best placed are those who learn to work with it early.

Why the worry is reasonable

Earlier waves of automation mostly affected physical and routine work. Generative AI can do parts of knowledge work: writing, summarizing, coding, analyzing documents and answering questions. That puts office jobs in scope in a way they were not before.

Some roles will shrink. Work that consists mainly of producing standard text, handling routine queries or moving data between systems is the most exposed. Pretending otherwise does not help the people in those roles.

Why “AI will replace everyone” is also wrong

Most jobs are a bundle of tasks. AI handles some of them well and others poorly.

It is good at drafting, summarizing, searching large amounts of text, writing routine code, classifying requests and spotting patterns in data.

It is weak at understanding context it has not been given, taking responsibility for decisions, dealing with people who are upset or uncertain, working out what a client actually needs, and knowing when its own answer is wrong.

So in most roles, AI takes over some tasks and leaves others. A developer writes less boilerplate and spends more time on design and review. A support agent handles fewer password resets and more complex cases. The job changes. Whether it shrinks depends on how much of it was the automatable part.

History offers some perspective. Spreadsheets, the internet and earlier automation each removed certain jobs, changed many others and created new ones. That pattern may repeat, but it is not guaranteed, and the transition is hard for the people whose work changes first.

Where AI is creating work

New work is appearing around AI itself:

  • building and integrating AI into business systems
  • preparing and governing the data AI depends on
  • testing AI output, monitoring it in production and fixing it when it goes wrong
  • security and compliance for AI systems
  • redesigning business processes to use AI sensibly

There is also more demand for people who combine domain knowledge with AI skills: a claims specialist who can design an AI-assisted claims process, or a tester who can evaluate an AI model.

What employees can do

  • Use AI in your current work. Learn what it does well and where it fails, in your own tasks. People who understand the tools are better placed than those who avoid them.
  • Move toward judgment and responsibility. The tasks that involve decisions, people and accountability are the hardest to automate.
  • Keep your domain knowledge sharp. Knowing the business, the customer and the rules is what makes AI output useful.
  • Learn to check AI output. The ability to spot a confident wrong answer is valuable in every role.

What employers can do

  • Start with the tasks. Map the work a team does, then mark which tasks AI should handle, which need a person to review the result, and which should stay with people entirely. The same map shows where staff need training and makes the change easier to explain.
  • Redesign the work first. Decide what people should do with the time AI frees up before making decisions about team size.
  • Retrain early. Give people time and support to learn new tools before their role changes.
  • Be open about plans. Uncertainty does more damage to morale than honest news.
  • Keep people accountable for decisions. Use AI to inform hiring, lending, medical or legal decisions, not to make them unchecked.

What to do next

If you manage a team, list the tasks it does in a typical week and mark the ones AI could reasonably do.

If you want help with that assessment, our AI and data team works with enterprises to bring AI into their processes in a way people can trust and work with.

Related

AI and Data

Book a 30-minute call
Contact

Tell us where delivery is stuck

A 30-minute call with someone who has run enterprise delivery. You leave with a clear view of what we would do, even if you don't hire us.

Office1700 Pacific Ave, Ste 3630
Dallas, Texas 75201
Delivery centreGurgaon, India
Dubai officeIRIS BAY, Dubai, UAE
Follow us