The question has changed. Two years ago it was "is this cheating?" Now it's "which parts should I hand over?" — because at this point most resumes have touched an AI tool somewhere, and hiring teams can usually tell which ones leaned on it too hard.
The honest answer: AI is genuinely good at parts of this job and genuinely bad at others. The failure mode isn't using it. It's using it for the wrong parts.
Where AI actually earns its keep
- Formatting and structure. Consistent tense across bullets, parallel structure, a section order recruiters expect. AI does this instantly and never gets tired of it.
- Rewording. If your bullet reads "responsible for managing the weekly report," a model will hand you five alternatives, and one of them will contain an actual verb. That's a real service.
- ATS compatibility. Knowing what parses cleanly — standard headers, no text buried in images, no clever multi-column layouts — is boring knowledge AI has in abundance.
- Pressure-testing. Ask it to find the weakest bullet on the page. It'll usually land on the same one you already suspected.
Where it quietly hurts you
- Voice. Models regress to the mean. Ask for "professional" and you get the same professional-sounding vocabulary everyone else got: dynamic, passionate, spearheaded, results-driven. Sameness reads as absence — like there's no specific person behind the document.
- Fabrication. Ask an AI to "improve" an achievements section and it will happily add numbers you never earned. It isn't lying maliciously; it's optimizing for what a good bullet looks like. But you're the one in the interview.
- Polish mistaken for substance. A beautifully written bullet full of mush is still mush. The formatting gets better; the case for hiring you doesn't.
The line worth drawing
Draft with it. Decide with yourself.
Draft with it. Decide with yourself.
The facts are yours — where you worked, what you actually did, what changed because you did it. The voice is yours — the difference between "drove alignment across stakeholders" and "got three teams who disliked each other to agree on a launch date" is the difference between a resume and a form letter.
Two tests. First, read the bullet out loud: if you'd never say those words to another human, don't submit them on paper. Second, the interview standard: never write anything you can't defend for two follow-up questions. AI doesn't know what you can defend. You do.
What this looks like in a tool
The design of the tool matters more than the fact that it uses AI. A one-click generator — paste a job description, receive a resume — puts the model in charge of the parts it's worst at. An editor that flags weak bullets, scores the draft against a specific posting, and leaves every change to you keeps it doing what it's best at.
Aptly is built the second way: you write, it marks up and scores, you decide what to take. That red-pen loop matters. The model never gets the final word on your work history.
Use AI the way you'd use a very fast, very literal-minded friend who's read a million resumes: great for structure, formatting, and a second opinion; wrong for the facts, the voice, and the judgment. Keep those, and the tool makes you sharper. Hand those over, and you get a document polished in exactly the same way as everyone else's.