The short answer
Yes, AI can help with your CV, and it is genuinely good at some parts of the job: tightening flabby sentences, turning vague duties into concrete achievements, matching your wording to a job advert, and catching the typo you have read past nine times. Used that way it saves hours and produces a better document.
What it cannot do is write your CV from nothing. If you hand it a job description and ask it to produce a candidate, it will invent a plausible one, and that plausible person is not you. Every fabricated line is something you will be asked about in an interview, and something an employer can treat as grounds for withdrawing an offer later. The rule is simple: AI edits your facts, it does not supply them.
What AI is genuinely good at
Turning duties into achievements. Most CVs list responsibilities: "responsible for social media." Hiring managers want outcomes: what changed because you were there. AI is good at interrogating you for that. Give it your bland bullet and ask what is missing, and it will ask about scale, results and timeframe. You supply the real answers.
Cutting the waffle. "Utilised a variety of methodologies in order to facilitate improved outcomes" becomes "cut order errors by rewriting the picking process." One prompt, ten seconds.
Tailoring to a specific advert. This is the highest value use. Paste your CV and the job description, and ask which of your existing experience matters most for this role and how to reorder it. You are not adding anything new, you are surfacing what is already true.
Fixing structure and consistency. Tense, date formats, verb style, the order of sections. Tedious, easy to get wrong, and exactly the kind of thing a machine handles well.
You are an experienced recruiter in [industry].
Here is my CV bullet: [paste one bullet].
Here is the job advert I am applying to: [paste it].
Ask me up to five questions that would let you turn this into a stronger, evidence-based bullet. Do not rewrite it yet. Do not invent any numbers or facts. Wait for my answers.
That "do not rewrite it yet" is doing a lot of work. It forces the tool to pull information out of you rather than making it up.
What gets you rejected
Fabrication. This is the line. Inventing a job, a qualification, a certification, a date range or a metric is not creative writing, it is lying on an application, and depending on where you are and what the role is, it can be grounds for dismissal years later. AI fabricates by default when it lacks information, and it does it in a confident, professional tone that reads beautifully. That is precisely what makes it dangerous. If you cannot defend a line for five minutes in an interview, cut it.
The generic AI voice. Recruiters read hundreds of applications and they have learned the tells: "leveraged synergies to drive impactful outcomes", "passionate about delivering excellence", relentless tricolons, an oddly formal register that does not match how you speak in the interview. A CV that reads as machine-assembled signals low effort, which is the opposite of what you want. Write in your own register and use AI to sharpen it.
Mismatch with your interview. If your CV claims fluency in something you touched once, the interview will find out in about four minutes. Awkward for you, and a wasted afternoon for them.
Losing the specifics. AI editing tends to smooth. Ask it to shorten a bullet and it will often drop the exact detail that made you interesting: the unusual system you used, the size of the budget, the fact that you did it alone. Read every edit and put back anything specific it quietly removed.
Never paste anything into a chatbot you would not be comfortable seeing outside your control. A CV contains your full name, address, phone number and employment history. Trim your personal details before pasting, and check the tool's data settings if you use it regularly.
What applicant tracking systems actually do
There is a lot of folklore here, so here is a level-headed version.
An applicant tracking system, or ATS, is the software an employer uses to receive and organise applications. Its main job is boring administration: storing your file, parsing it into fields (name, employer, dates, skills), and letting a recruiter search and filter the pile. Most systems do not score you out of a hundred and reject you automatically. What usually happens is that a human searches for particular keywords or filters on a stated requirement, and applications that do not surface in that search never get read.
The practical consequences are less dramatic than the internet suggests, but they are real.
Use the words from the advert, where they are true of you. If the role says "accounts payable" and your CV says "supplier payments", you may not appear in the search. This is not gaming the system, it is translation. If it is not true of you, leave it out.
Keep the file simple. Standard headings ("Experience", "Education", "Skills"), a normal font, no text inside images or graphics, no essential information in headers, footers or sidebars, no multi-column layouts that scramble when parsed into plain text. Send a Word document or PDF as requested, and if no format is specified, either is usually fine.
Write dates plainly and consistently, in a format like "March 2023 to July 2025". Parsers are better than they used to be, but they still get confused by creative formatting.
Ignore anyone selling you invisible white text stuffed with keywords. It is easily detected, it looks like deliberate deception, and it will get your application binned by the human who spots it.
A quick test: save your CV, then copy all the text out of it and paste it into a plain text editor. What you see is roughly what the parser sees. If it comes out scrambled, out of order, or missing a section, your layout is the problem.
A workflow that is both honest and effective
- Write the raw material yourself. Every role, every real achievement, every number you can actually stand behind. Ugly is fine at this stage.
- Ask AI to interview you. Use the prompt above, one bullet at a time. Answer its questions truthfully. This is where the value is.
- Tailor for one specific role. Paste the advert and ask which of your experience matters most and what to lead with. Do not add anything new.
- Ask for a critique, not a rewrite. "You are a hiring manager for this role. What are the three weakest parts of this CV and what would you ask me about?"
- Read it aloud. Anything that does not sound like you, rewrite in your own words. This single step removes most of the machine tell.
- Verify every fact. Dates, job titles, figures, qualifications. If AI touched a line, check the line.
Used this way, AI is a strong editor and a useful interviewer. It is not a candidate, and the moment you let it become one, you have created a problem for a version of yourself who is sitting in an interview room with no idea what is on the page.
If you want the whole process, including cover letters, tailoring at speed, interview preparation and how to keep your applications honest, our free AI-Powered Job Search course walks through it step by step.