White Papers
ChatGPT Made Your Posting Sound Better. It Still Lost You the Person You Actually Wanted.
Why generic AI is a wider net — not a better one.
3
choices every company has when writing a job posting — only one of them works
0
new information added by ChatGPT — it can only polish what was already on the page
5 decades
of recruitment marketing experience behind JobWriter's behavioral approach
01
The promise that wasn't kept
Here’s what happened before AI: your postings got plenty of responses, but most of it was noise you had to wade through to find anyone worth a second look.
The people you wanted came through sometimes — never reliably, and never without costing you the time spent sorting through everyone else. Then ChatGPT and other AI tools showed up looking like the fix. You ran your job descriptions through it. The quality of what came back did improve. You told yourself the problem was solved.
It wasn't. And here's the part that should sting: ChatGPT doesn't filter out the wrong people — it can't, it has no way to know who's actually wrong for this specific role.
What it does is make the posting friendlier, and friendlier invites more of everyone, good fits and bad fits alike, in ten seconds, for free, without knowing anything about your team. That's not filtering. That's a wider net. The improvement was real — more people responded. But the improvement that mattered never came, and the noise you have to sort through may have grown, not shrunk.
02
What it actually cost you
Here’s what that cost you, and it’s not hypothetical: the exact person you wanted read the nicer posting, felt nothing specific, and took an offer from a company that happened to say the one thing that made her feel like this particular role was written for her. You will never know that happened. Your response numbers went up. It looked like it worked.
Here’s why: nothing new got said. The specific reason the exact right person would have loved your particular job — the detail that would have made her stop scrolling — was never in the original posting, because nobody had articulated it yet. ChatGPT can’t add information nobody gave it. It can only take what was already there and say it more pleasantly. That’s not an OpenAI problem — it’s true of any AI tool used the same way.
That missing piece is usually tacit knowledge sitting in a hiring manager’s head, never written down — and it’s a well-documented bias that people with deep, specific knowledge consistently fail to translate it into language a stranger would need, even when it’s directly in their interest to do so.
Camerer, Loewenstein & Weber, 1989
doi.org/10.1086/261651Generic AI doesn’t have access to that knowledge either. It was never given it. It can only polish what made it onto the page in the first place — and if the real hook never made it onto the page, there’s nothing to polish into something that works.
03
Why the gap goes unnoticed
Taking any action against a perceived problem lowers a person's sense of risk, which reduces motivation to look for a better fix — even when the action taken barely addressed the real issue. A better response satisfies that instinct completely, which is exactly why almost nobody checks whether it's the best one available.
Weber, 2006
doi.org/10.1007/s10584-006-9060-3ChatGPT
The Horse
A real improvement over walking. Gets you there faster. But it's not the car — and choosing it means you never find out how much faster the second one was.
The car
JobWriter
A different category of solution entirely. Not the same trip at a different speed — a fundamentally different way of reaching the right person.
04
Three choices, one answer
Every company writing a job posting has the same three choices:
Say the bland thing everyone already assumes
Say a nicer version of the same bland thing
Say the specific, true thing
the right person needs to hear at the one moment they're paying full attention. Only the third one was ever going to change who applies.
05
What real specificity promises
Be clear about what that third option promises, because overclaiming it costs more than it’s worth. This isn’t a rainmaker. It won’t hand you a perfect match or an overnight turnaround. What real specificity does is simpler and more honest: it improves your odds of getting more of the right people, more consistently, than either of the other two choices ever could. That’s a real gain. It’s not a guarantee, and nobody should sell it as one.
You can do this yourself. Ask the hiring manager what makes this specific role, this specific team, different — not “great culture,” the real, specific thing. Write that down before anyone writes a single word of the posting — then write the posting the way you’d talk to that one person, not a crowd. You’re asking someone to consider changing their life. That deserves a direct, human tone, not a corporate one.
That’s the whole method, and it works whether or not you ever use JobWriter. But if this is the exact problem you’re reading this to solve — a posting that sounds better and still isn’t landing — this is exactly what JobWriter was built to do, surfacing exactly the kind of detail a hiring manager wouldn’t think to volunteer on their own, every time, not a related idea. The actual answer.
See how JobWriter approaches this differently.
Role-specific behavioral targeting — built into every posting, by default.
Request a Demo