AI in employer branding - AI for scale, people for substance

AI in employer branding is no longer a question about the future. Almost every employer with a hiring problem now asks the same three things. Can AI write our job ads? Can we generate employee stories instead of filming them? And should we be doing something about the AI assistants candidates ask their questions to? Underneath all three sits one real question: will this save us time and money, and what does it cost us if we do it?

The honest answer is that AI delivers enormous value in some places and does the exact opposite of what you want in others. The line between them is not complicated, but you have to draw it before you start. We draw it in the same place every time: AI for scale, people for substance.

Where does AI in employer branding actually work?

Anywhere the job involves scale, repetition and speed. Ad variants per audience, per region, per platform. Where you used to test three versions, you now test twenty, and that is not a luxury. It is the difference between guessing and knowing. The same goes for subtitles, new cuts of existing video, formats for five different placements and translations into Polish or Romanian for audiences you would otherwise never reach.

It is strong in analysis too. Combing through application data, exit interviews, reviews and open answers from an employee survey takes a person days. A model does it in an afternoon, which leaves you time to test what comes out with the people it is about.

The common thread: AI is good the moment the original is right. It multiplies what you already have. It does not invent what is true.

Why do candidates spot an AI-written job ad so easily?

Because a language model produces the average of everything that already exists. In recruitment, that average is precisely the problem. Driven team, ambitious organisation, we are looking for a real go-getter. Every employer says it, so it says nothing.

A technician weighing up a move wants to know very different things. Which van does he drive, how far is the first job, how many evening shifts are in a month, who will his foreman be and what happens when he wants to train for something new in two years. None of that lives in the average, because it belongs to you alone.

And trust is thin. Research by Greenhouse among more than 4,000 jobseekers in the US, UK, Ireland and Germany found that 45 per cent of British candidates trusted hiring processes less than they had a year earlier, and that 40 per cent blamed the growing use of AI. Visibly letting AI write for you carries a price.

What does it mean that candidates ask ChatGPT about you first?

It means you are no longer the only one telling your story. A study by Gdoc.io among 500 American jobseekers found that 36 per cent use AI in their search, and that nearly 80 per cent of them use ChatGPT for it. A clear majority never mention it to the employer.

In May 2026 PerceptionX surveyed more than 300 candidates across seven countries, almost all of them AI users. The topics they asked about most: pay, career progression and what the hiring process is like. Close to six in ten had at some point been given incorrect information about an employer.

That is the point. You are being described whether you take part or not. If your own careers site says nothing concrete about pay, rosters and progression, a model fills those gaps with job boards, review sites and old press coverage.

How do you get AI in employer branding to retell your story properly?

You cannot steer a model, but you can feed it. What is stated consistently, explicitly and in more than one place is what gets picked up. What you merely imply disappears.

In practice that means putting your salary range, shift patterns, travel distance, training budget, the hiring process step by step and the progression routes into plain sentences on your own pages. Not a pdf, not text inside an image, not a vague clause where a number belongs. And make sure it holds up off your own site as well, because reviews and job boards count.

The useful part is that none of this asks you to do anything you should not be doing anyway. Everything you write down so a model retells it correctly is exactly what a hesitant tradesperson needs in order to decide. This is not an AI trick. It is writing your employee value proposition down in the place where it actually gets read.

Where does AI in the recruitment funnel become risky?

In screening and matching. The European AI Act treats AI systems used for recruitment and selection as high risk. The heaviest obligations were pushed back during 2026 to the end of 2027, but transparency rules land earlier and the direction of travel is clear.

More important than the date is what automated filtering does to who comes in. In blue collar sectors a lot of good talent sits behind a CV a model does not recognise. Career changers, people with a patchy work history, craftspeople whose quality lives in their hands rather than on paper. In that same Greenhouse research, only 8 per cent of candidates believed AI makes hiring fairer.

Filtering is fine, and it is necessary. An honest picture of the work makes the wrong candidates drop out, and that is a gain. But then you filter on what you decided matters, not on what a model happens to recognise.

Why fake a person you could simply film?

This is the sharpest line of all. You can have an avatar play an employee, have a model write a testimonial and render a shop floor that never existed. All of it is possible. And all of it is a waste, because the one thing no model can produce is the sentence your technician comes up with himself when you ask him why he is still there after eleven years.

That is not just a matter of taste. Lighthouse Research & Advisory found that 55 per cent of active jobseekers consider employee video more credible than company produced video, and that a video from the hiring manager lifted the likelihood of candidates considering the role by 46 per cent. People want to see who they will be standing next to.

It is how we work. For Hanab, part of VolkerWessels, we make content where the work happens, with the people doing it. At Picnic the story is about the route someone can take through the organisation, including the Picnic Academy. You do not claim a culture, you let people feel it, and that works best when the shop floor puts it into words rather than the board.

What does AI in employer branding not change?

The question underneath. Vacancies that fail to fill are not a problem, they are a symptom. Sometimes too few people apply, sometimes the wrong ones do, sometimes time to hire is the leak and sometimes a third of new starters leave within a year. Those are four different solutions, and no model picks one for you.

So it starts with a baseline. Which number has to move, where does it stand today, and when do we call it a success? Agree that up front. Then you can show afterwards what it delivered, which also helps whoever has to defend the budget internally.

AI makes a good approach faster and a wrong approach faster in the wrong direction. And not every question is a campaign. Sometimes the answer is a shorter application process, a better career platform, or employer branding that stays visible all year instead of for six weeks.

How does MonFri approach this?

We start with your question. Which number has to move, and why is that not happening now? From there we look at the three pillars under your employee value proposition: what the outside world says about you, what your own people say, and what the market offers. The distinctiveness sits in the overlap, and that overlap is different for every employer.

Only then do we talk about means. We use AI heavily for variants, tests, analysis and speed of rollout. The substance we collect from your people, with a camera, on location. That is not a principle, it is arithmetic: it simply works better.

Want to know what that looks like for your organisation? Have a look at our services, call 020 23 38 587 or email hi@mon-fri.com. We will start with the question before we talk about a campaign.