What Really Happens to Your Application.

 

Let’s start by getting rid of something. You’ve almost certainly seen the claim that 75% of CVs are rejected by a machine before a human ever sees them. It’s on careers blogs, LinkedIn posts and half the CV-writing services online.

There’s no credible source for it. Trace it back, and you arrive at a marketing claim made by a recruitment software vendor in 2012 – a company that no longer exists. It’s been recycled ever since because it’s frightening and it sells things.

So, we’ll let this go. The reality is less alarming and more useful.

 

What’s actually happening

Most large employers use an Applicant Tracking System. Its main job is administrative: it takes your CV apart and files the contents into structured fields – name, dates, job titles, qualifications, skills – so a recruiter can search and sort a thousand applications without opening a thousand documents. The University of Bristol’s careers service, which is refreshingly blunt on this, says that automatic rejection based on CV content alone is “very unlikely.”

AI is a separate question, and it’s growing. The Chartered Institute of Personnel and Development (CIPD)’s most recent Resourcing and Talent Planning survey found around 31% of UK organisations using some form of AI or machine learning somewhere in recruitment, up from 16% two years earlier. Worth reading that carefully, though: it’s AI anywhere in the process. The proportion using AI specifically to shortlist candidates against a job description was a lot smaller – around one in twenty.

So: you are unlikely to be rejected by a robot. You are quite likely to be sorted, ranked and searched by software before a human decides where to look first. Those are different problems, and the second one you can do something about.

 

Why good candidates still get filtered out

Harvard Business School’s Hidden Workers research asked employers directly about this, and their answers were striking. A large majority – 88% – agreed that qualified candidates were being screened out because they didn’t match the exact criteria in the job description. Not because they couldn’t do the job. But because the system was looking for a precise phrase and found a different one.

That’s the real risk, and it’s a translation problem more than a technology problem.

 

What actually helps

Use their words (honestly!) If the ad says “stakeholder management” and you’ve written “dealing with clients and suppliers”, change it – provided it’s true. Stuffing your application with keywords doesn’t work and makes you harder to read, but genuine alignment of vocabulary does.

Keep the layout plain. Single column. Standard headings – Experience, Education, Skills. Skip the columns, text boxes, icons, graphics and creative header banners. Anything visually clever risks being parsed (converted by the computer) into nonsense, and the human reading it afterwards sees the nonsense, not your design.

Spell things out both ways. On your first mention, write the full term and the abbreviation together, e.g. “The Chartered Institute of Personnel and Development (CIPD)”. You don’t know which one the system is searching for.

Pro tip: Send a Word document unless the employer specifically asks for a PDF. Word remains the more reliably parsed of the two.

 

And the other direction

It’s worth knowing employers are now watching AI use by candidates too. The Institute of Student Employers found that a third have redesigned their selection process in response, and 61% reported candidates using AI during interviews without saying so. Most aren’t opposed in principle – 61% are fine with AI helping write a CV or covering letter. What they object to is undisclosed use when they’re trying to assess you.

There’s a fairness argument being had, too. The Information Commissioner’s Office audited AI recruitment tools and found some inferring candidates’ gender and ethnicity from their names, issuing nearly 300 recommendations as a result.

None of this is a reason to despair about applications. It’s a reason to make yours easy to read – by both kinds of reader.

What’s one phrase on your CV that doesn’t match the language of the jobs you’re applying for?