If you've applied to jobs at any company over 200 employees in the last decade, your resume probably went through an Applicant Tracking System (ATS) before any human read it. ATS software ranks resumes by keyword match against the job posting. Resumes that don't match get filtered out before recruiters scan them. AI is genuinely useful for fixing this — but the trap is producing a resume that's keyword-stuffed, generic, and still doesn't get you interviews.
Two real problems AI helps with
Problem 1: keyword gap. Your resume describes what you did using your own words. The job posting describes what they want using their words. ATS scoring reads the difference as poor fit. AI can rewrite your accomplishments using their vocabulary without distorting what you actually did.
Problem 2: weak phrasing. Most people write "responsible for managing the budget" instead of "reduced budget by 22% while delivering the same outcomes." AI is excellent at converting vague duties into specific, quantified achievements when you give it the underlying facts.
What AI shouldn't do
- Invent achievements that didn't happen
- Add skills you don't have
- Inflate titles you didn't hold
- Make up numbers when you don't know the actual ones
This isn't just ethical. ATS systems and recruiters increasingly cross-check claims against LinkedIn and references. Caught lies kill candidacies and reputations. Use AI to express your real experience better, not to fabricate it.
A workable workflow
Step 1: Keep a master resume document with everything. Every job, every accomplishment, every project — long-form, with real numbers (revenue, headcount, percentages, timeframes). This is your truth source. AI works on this; you don't apply with this.
Step 2: For each application, paste the job description into Claude or GPT. Ask: "Extract the top 15 skills, technologies, and qualifications mentioned in this posting, ranked by emphasis."
Step 3: Match against your master resume. Ask: "Looking at my master resume below, which of those 15 skills can I credibly claim, and which accomplishments demonstrate each?"
Step 4: Generate the application-specific resume. Ask: "Rewrite my resume tailored to this posting. Use vocabulary from the posting. Lead each role's bullets with the most relevant accomplishments. Keep all claims truthful — don't invent or inflate. Use specific numbers from my master where available."
Step 5: Edit by hand. AI will produce something usable but slightly generic. Read it through and rewrite anywhere it sounds like AI. Trust your ear.
Common AI resume mistakes
- Buzzword soup. AI defaults to filler like "strategic," "results-driven," "passionate," "detail-oriented." These add nothing — they show up on every resume. Cut them.
- Generic action verbs. "Led," "managed," "oversaw" repeated. AI suggests these because they're safe. Replace with specific verbs: "shipped," "closed," "hired," "reduced," "automated."
- Bullets that don't quantify. "Improved customer experience" is meaningless. "Reduced support tickets 31% by rewriting onboarding flow" is what gets you a callback.
- Skills section bloat. AI adds every plausible skill. Recruiters skim the skills section in 3 seconds. Keep it to skills you can defend in an interview.
- Tense inconsistency. Past tense for past jobs, present tense for current. AI sometimes mixes.
Format: ATS-friendly is genuinely friendly
The industry has moved away from creative resume formats. ATS still struggles with:
- Multi-column layouts (parsing breaks)
- Tables (parsing breaks)
- Graphics or icons (ignored or break parser)
- Headers and footers (often skipped entirely)
- Non-standard section names ("My Story" instead of "Experience")
Use a single-column, plain-text-friendly format with standard section names: Summary, Experience, Skills, Education. Save as both PDF and DOCX; some systems prefer one or the other.
Visual design is for humans. ATS sees plain text. Get past ATS first; then your design choices reach humans.
When NOT to use AI for your resume
For academic CVs, AI tends to flatten the precise format conventions academic search committees expect. If you're applying for tenure-track or research positions, follow the conventions of your field rather than letting AI generic-ize.
For highly visual fields (designers, illustrators, art directors), the resume isn't the main artifact — the portfolio is. AI rewriting a designer's resume helps marginally; investing the time in portfolio quality helps more.
For applications via personal connection. If a friend referred you, the resume is half the equation. AI keyword-optimization doesn't help the conversation that follows. Be authentic; be yourself.
The cover letter question
AI is even more useful for cover letters than resumes — they're more flexible, lower-stakes, and easier to personalize. The trap: AI-written cover letters are detectable. Recruiters who read 50/day spot the pattern.
If you use AI for cover letters: use it for structure and first draft, then rewrite at least the opening paragraph and the closing in your own voice. The middle (specific examples linking your background to the role) can be AI-drafted but should be substantively yours.
What about LinkedIn
The same workflow applies to LinkedIn profiles. Bigger considerations:
- Recruiters search LinkedIn with similar keyword logic to ATS
- Profile photo and headline matter for click-through
- Featured projects (if relevant) substantially improve impressions
- AI can help write a strong "About" section that includes both keywords and personality
Don't let AI write your LinkedIn story. The story sections are where personality differentiates. Keywords elsewhere; voice in story.
Decision tree
- Applying to large company with ATS: AI for keyword optimization, hand-edit for voice
- Personal connection / referral: be yourself, light AI polish only
- Academic / research: follow field conventions, AI helps marginally
- Creative field with portfolio: portfolio matters more than resume
- LinkedIn: AI for keyword sections, never for personal story
Next steps
- Build the master resume document; keep it updated
- Read about ATS systems specifically (Greenhouse, Lever, Workday differ)
- Track which applications get callbacks; A/B test resume variants
- Use AI to generate interview prep questions based on the job description