How to Write a Resume With AI

Okay so I want to start by saying I was skeptical. I spent like two years tweaking my resume word by word, agonizing over whether to say "managed" or "led" or "coordinated," and the idea of handing that over to an AI felt wrong. Like cheating but also lazy? Both.

And then I got laid off and needed to apply to 30 jobs in two weeks and I got over it really fast.

Here's what I learned from doing this a lot in a short amount of time. AI is genuinely useful for resume writing. But not in the way most people think. It's not a magic resume generator you dump your life into and out pops a perfect document. That approach produces garbage. I know because I tried it first.

The wrong way (so you don't do what I did)

My first attempt: I pasted my old resume into Claude and said something like "improve this resume." What came back was technically fine but felt completely hollow. Every bullet point was suddenly full of words like "spearheaded" and "orchestrated" and it read like it was written by someone who had never met me. Hiring managers can smell that from miles away and I knew it immediately.

The problem is that "improve my resume" is an impossible instruction. The AI doesn't know what job you're going for. It doesn't know what you actually want to emphasize. It doesn't know which of your accomplishments you're proudest of. It just takes what you give it and makes it sound more... resume-y. Which is not what you want.

The approach that actually worked

What works is using AI as a thinking partner, not a ghostwriter. Here's the actual process I settled into.

First, I'd find a job posting I wanted to apply to and paste the whole thing into my conversation. Not a summary of it, the whole thing. Then I'd paste my existing resume. Then I'd say something like: "I'm applying for this specific job. Based on this posting, what experience in my background should I be emphasizing most? What's missing? What keywords are they clearly looking for?"

That's a completely different question and you get completely different answers. The AI would come back with things like "they mention 'cross-functional collaboration' three times, you have that experience but your current resume buries it in one bullet" or "they're looking for someone with direct client communication experience and you don't mention your freelance work anywhere." That's actually useful. That's a smart friend who read the job description carefully.

Then I'd write the bullets myself, based on what I learned from that conversation. And then I'd ask the AI to help me tighten the language. That order matters. Write it yourself first, then polish. Not the other way around.

The keyword thing is real and kind of annoying

I'm not going to lie, the keyword optimization part feels a little gross. Like you're gaming a system. But the system is real, most companies use applicant tracking software that literally filters resumes before a human sees them, and if your resume doesn't include the specific words from the job posting, it gets filtered out. I've confirmed this with multiple recruiters who told me directly.

So I got over my feelings about it. After I had a draft I was happy with, I'd ask Claude to compare my resume against the job posting and tell me which key phrases I was missing. Then I'd figure out if I could honestly work those phrases in somewhere. If I couldn't honestly include something because I didn't have that experience, I didn't add it. I'm not suggesting you lie. But there's usually a real version of what they're asking for somewhere in your background and sometimes you just need help seeing it.

The summary section is where AI helps most

Okay here's my hot take: the summary/objective section at the top of a resume is both the most important part and the part people write worst. Including me, historically.

I used to have some generic three-sentence thing about being "a results-driven professional" with "strong communication skills" that said absolutely nothing. After going through this process I realized the summary should basically be a cover letter compressed into two sentences, specifically why you're the right person for THIS job, not a generic description of yourself as an employee.

For this part I actually found AI super helpful. I'd describe the job I was going for, describe my background honestly, and ask it to help me write a summary that connected the two clearly and specifically. I'd edit it heavily after, but having something concrete to react to was way faster than starting from scratch.

A thing I wish someone had told me

Don't ask AI to write your whole resume and then just... send it. Read every single line out loud. I mean literally out loud.

Here's why. AI occasionally generates bullet points that sound plausible but are slightly off from what you actually did. Like it'll say you "managed a team of six" when actually you coordinated with a team of six, different thing, and one of them is potentially a lie you'd have to defend in an interview. If you haven't read every line carefully, you might not notice until someone asks you to elaborate and you're like "wait did I say that?"

I caught two things in my own resume that I would've been embarrassed to defend. Read it out loud. Every word.

Which AI to use

Honestly, Claude, ChatGPT, and Gemini all work fine for this. I've used all three. Claude tends to give me feedback that feels a little more nuanced when I'm asking analytical questions like "what's missing from this resume." ChatGPT is maybe slightly better at punchy language if I need help tightening a bullet. Gemini integrates with Google Docs which is convenient if that's where you're working.

But I'm not going to tell you one is definitively better because I think the difference is pretty small and what matters more is how you're prompting them. The framework I described above works across all of them.

What I'd say is: don't pay for a resume writing service that charges you hundreds of dollars to do exactly what you can do yourself with a free AI account and an afternoon. I've seen what those services produce. It's the same stuff.

The part that's still on you

AI cannot tell you what you want. It can't figure out which version of your career story is the most compelling one. It can't decide whether to lead with your technical skills or your management experience. Those are judgment calls that require you to have a point of view about your own career.

The people who get the most out of AI for resume writing are the ones who come in with opinions and use the AI to execute and refine. The people who get the least out of it are treating it like a vending machine where you put in your job history and out comes a resume. It doesn't work that way.

Come in with thoughts. Let the AI push back. Rewrite. Repeat. That's the actual process and it works really well once you stop expecting magic and start expecting a good collaborator.

I went from getting zero callbacks to getting five in one week. Same experience, totally different resume. Worth the afternoon.