AI for job search: from the job ad to the interview
GPT can break down a job ad, tighten your CV bullets and play a demanding interviewer. It only knows what you tell it, so it must never add experience you do not have. Here is a workflow that stays honest.

Job hunting involves a lot of writing that is hard to do well about yourself: CV bullets, cover letters, short answers on application forms, follow-up notes. GPT is good at exactly this kind of text, and it is a patient interview partner too. It also knows nothing about you except what you type, and it will happily make your experience sound bigger than it was if you let it. This guide walks through a job search with our free GPT chat, step by step, in a way that keeps every claim true.
Read the job ad properly
Job ads mix real requirements with wish lists and company boilerplate. Before you write anything, ask the model to sort them.
Here is a job ad. Sort its contents into: 1. Must-have requirements (deal-breakers) 2. Nice-to-have skills 3. What the daily work will actually involve 4. Words and phrases they repeat, which I should mirror 5. Questions I should ask in the interview Job ad: [paste the full ad]
This takes a minute and changes how you apply. You see which of your experiences matter most, which keywords an applicant tracking system may look for, and whether the role is what the title suggests. Open this prompt in the chat and paste the ad below it.
Tailor your CV without inventing anything
The goal is to present your real experience in the language of this particular job. Give the model your current CV and the analysis from the previous step, and set firm rules.
Weak: Make my CV perfect for this job. Strong: Here is my CV and the key requirements of a job I'm applying for. Rewrite my last two roles' bullet points so the most relevant experience comes first and uses the ad's wording where it is accurate. Rules: do not add any skill, tool, number or result that is not in my CV. If a requirement isn't covered, list it separately so I can decide what to do.
That last rule matters. A model trying to be helpful may turn "helped with the newsletter" into "led a marketing campaign that grew subscribers". Read every changed line and ask yourself whether you could defend it in an interview. If not, change it back.
Good CV bullets show what you did and what changed because of it. If you have real numbers, such as hours saved or orders handled, add them yourself; the model must not guess them.
Cover letters that sound like you
Generic cover letters are easy to spot, and AI makes them easier to produce. To avoid that, bring the specifics yourself:
- one reason you want this company, in your own words;
- one or two stories from your work that match their needs;
- a sample of how you write, such as an email you were happy with.
Then ask for a short draft, around 250 words, that uses those points and your style, and rewrite it by hand. Delete any sentence that could appear in anyone's letter. Our writing tips list phrases to cut, and the email writer helps with the short message you send with your application.
Practise the interview
This is where a chat helps most, because rehearsing out loud with a friend is awkward and a model does not get bored. Set up a role-play:
Act as the hiring manager for this role: [paste job ad]. Interview me with eight questions, one at a time: a mix of behavioural, technical and motivation questions. After each of my answers, give two lines of feedback: what worked, and what a stronger answer would include. Be demanding but fair.
Answer in full sentences, as you would in the room. For behavioural questions, many interviewers expect a short story with a situation, the task, your action and the result, often called the STAR method. Ask the model to point out when one of those parts is missing.
When you finish, ask: "Which of my answers was weakest, and what follow-up question would a real interviewer ask about it?" Then practise that one again. If you want a tougher interviewer, press Think to use GPT-6.1 Terra, which takes a moment to consider your answer before it replies.
Salary, offers and hard conversations
GPT can help you prepare for a salary conversation: list arguments based on your experience, rehearse phrases for asking for more time, or draft a polite reply to an offer. It cannot tell you the going rate in your city right now. The chat cannot go online, and the models learned from data that ends at a cut-off date, so any salary figure it offers is a rough guess from older material. Use salary surveys, job boards and people in your field for numbers.
The same goes for employment law, notice periods and contract terms: the model can explain general concepts and help you write questions, but rules differ by country, so check with an official source or an adviser before you rely on anything.
What to keep out of the chat
A CV is full of personal data. Before you paste it, remove what the task does not need:
- your full address, phone number and personal e-mail address;
- date of birth, ID or passport numbers, national insurance or social security numbers;
- references' names and contact details;
- anything confidential about your current employer, such as unreleased plans or client names.
Replace them with placeholders like [PHONE] and add them back in your own copy. Our privacy and safety guide explains what happens to your messages on gpt.me and how to redact quickly.
Follow-ups and thank-you notes
After an interview, a short, specific note helps you stand out. Tell the model what you discussed and one thing you want to add, and ask for a message of about 80 words. After a rejection, it can help you write a gracious reply that asks for feedback, which keeps the door open. The career prompts have ready-made versions of these, and the assistants include helpers for CVs and interviews.
Keep each application in its own chat. Your history stays in your browser, so you can return to a company's chat before the second interview and pick up where you left off, and you can export a chat if you want a copy.
FAQ
Can employers tell my CV was written with AI?
They notice generic wording and claims that do not hold up in conversation. A CV built from your real experience and edited by you reads like yours, however you drafted it.
Is it honest to use AI in a job application?
Using it to organise and phrase your real experience is similar to asking a friend to proofread. Adding skills or results you do not have is not, whoever writes it. Some employers state rules on AI use; follow them.
Can GPT find job openings for me?
No. Our chat cannot search the internet or job boards. It works with the ads you paste. Use job sites for searching and the chat for reading and writing.
Which model should I use?
The default GPT-6 Luna is fast and fine for CV bullets, letters and notes. For a demanding mock interview or for comparing two offers point by point, try GPT-6.1 Terra with the Think chip.
Got a question? Ask it now
No account and no download. Open the chat, type what is on your mind, and the answer starts appearing in a second or two.
Start chatting

