When Every Application Is Well Written

For a long time a cover letter told you something, and it's worth being precise about why. The words themselves weren't that revealing. What mattered was that writing a good one took time. Someone who spent an hour learning about your company and explaining why they wanted the job had shown you, just by spending the hour, that they wanted this job more than the others they could have applied to.

Most written screening rests on the same idea. Application forms, knockout questions, cover letter analysis, even most AI resume parsers assume that the quality of what a candidate writes reflects something real about them. A careful answer to "Describe a challenge you overcame" suggested a person who could think a problem through. That assumption held up as long as writing the answer was hard.

It isn't hard anymore. Anyone can paste your job description into a modern AI model and get back, in under a minute, a cover letter that is specific, tailored and grammatically clean, and that reads like one somebody labored over for an hour. The same is true of application questions, skills assessments given as text prompts, and follow-up emails. Lots of candidates now use AI to write, tailor and bulk-submit their applications. I wouldn't call that cheating. They're using the tools in front of them, as you would. But it means the written part of your funnel now mostly measures whether someone can use AI, and everyone can.

What happens to a screening process built on writing when writing is free? Volume goes up, because each application costs so little to produce, while the quality of the writing converges on the same competent floor. Everyone looks adequate on paper and almost no one stands out. A recruiter ends up reading slightly different versions of one application hundreds of times.

What's worse is that neither pile can be trusted. The rejects may include someone strong whose application happened to look like everyone else's. Meanwhile the shortlist may include people whose main talent was a good prompt. Strong written answers used to predict strong candidates, and now they mostly don't.

The natural response is to score resumes faster, with more automation. I don't think that helps. If the text going in no longer says much about the person, processing it more efficiently only gets you to the same unreliable answer sooner. Your ATS and screening tools are working as designed. They're just moving data that has lost most of its meaning.

If the old filter worked because it cost the candidate some effort, the question to ask is what still costs effort. I think the answer is a conversation. A candidate can produce a flawless written answer in ninety seconds. Nobody can produce a flawless live phone screen on demand, least of all one where the next question depends on what they just said. What comes out in a conversation is how the person actually thinks, how they explain things, and how much they really know about the work.

Imagine two applicants for the same role. One wrote a perfect cover letter but stumbles when asked about the basics of the job. The other sent a generic application but talks clearly and specifically about work they've really done. A text screen would probably advance the first and reject the second. A few minutes of conversation sorts them the right way round.

That reasoning is why we built Asendia to screen by talking. It's software that phones applicants and interviews them, and each follow-up question depends on the answer before it, so a good prompt doesn't get anyone very far.

Changing how you screen is the main thing, but I'd change two smaller things with it. One is the intake form. If your ATS still opens with ten written questions, you are asking candidates to generate AI-polished answers before any human sees them, and you'll learn little from what comes back. Cut it down to the basics and let the conversation do the qualifying.

The other is how recruiters decide whom to reject. Treating writing quality as a proxy for effort made sense in 2019. Now it works against you. A recruiter who reads a vague cover letter as low motivation is probably discarding someone who was simply moving fast, like every other applicant.

None of this looks temporary to me. Candidates have found that AI makes job hunting much more efficient, and they have no reason to give that up. So I'd treat a flood of AI-written applications as the normal state of things and build around it, instead of waiting for it to pass. If you're also rethinking what to measure now that AI is in the funnel, we wrote about recruitment KPIs in a post-AI world, which is the measurement side of the same problem.

A simple check is to look at what your first screening step actually relies on. If it's mostly what candidates wrote, it relies on the one thing that just became cheap.

Ready to transform your hiring strategy? Schedule a Demo with our founders today!