AI in recruitment: what it should — and shouldn’t — decide

A year ago, most hiring managers we spoke to were curious about AI. Today, nearly all of them are using it somewhere — to write job descriptions, to screen CVs, or to schedule interviews. The question has moved from “should we?” to “where should we stop?”

Where automation earns its place

Volume is where AI shines. When a new plant needs 60 operators or a GCC opens 40 engineering seats, first-pass screening against clear, objective criteria — certifications, years of experience, location — is exactly the kind of repetitive work software does well.

• Parsing and de-duplicating CVs from multiple sources
• Matching hard requirements such as licences or shift availability
• Scheduling interviews across busy calendars

“Software can tell you who meets the brief. It can’t tell you who will make the team better.”

Where judgement still wins

For senior and specialist roles, the signal lives in things a model can’t see: how a candidate talks about a failed project, whether their ambitions match the company’s next three years, how they’ll get on with a particular plant head. These are the calls our consultants make every day.

There’s also fairness to consider. Models trained on historic hiring data can quietly reproduce old biases. Any automated filter should be audited, explainable and — above all — reviewed by a person before it rejects anyone.

What this means for you

If you’re hiring, use AI to go faster on the obvious and spend the time you save on the conversations that matter. If you’re a candidate, make sure the basics on your CV are machine-readable — and then make sure a person hears your story.