How to use AI in your job search: what helps, what backfires and where to draw the line

Reading and triaging job ads

A 900-word ad can be turned into a short list of must-haves, nice-to-haves and deal-breakers in seconds: location rules, visa requirements, languages, years of experience. That alone saves time on roles you were never going to get. For remote roles, the free remote eligibility checker does a narrower version of this without any AI: it quotes the sentences that decide whether a “remote” job is open to where you live.

Comparing your CV with a job

AI is good at lining up two documents: which requirements your CV supports, which it doesn’t mention, and which of your achievements are most relevant. The free CV score checker gives you a 0–100 match and the keywords your CV is missing.

First drafts and editing

A first draft from your own notes, such as a cover letter skeleton, three versions of a summary or a rough bullet turned into a clean one, gets you past the blank page. AI is also a patient editor: tightening long sentences, fixing grammar, making tenses consistent. If you write in a second language, it helps with phrasing, though a fluent human should still read anything important.

Interview practice

Generating likely questions from a job ad, running a mock interview and getting feedback on the structure of your answers are among the most useful things AI does. Pair it with our STAR method guide so you know what a strong answer looks like.

Staying organised

Tracking where you applied, when to follow up and what you said is tedious. Templates help here even without AI: the free follow-up email generator drafts a polite check-in dated from when you applied.

• Invented facts. Language models can produce plausible details that aren’t true: a metric you never hit, a certification you don’t hold, a tool you’ve never used. On a CV that isn’t a typo; it’s a false claim with your name on it.
• A generic voice. Unedited output sounds alike across candidates. The same openers and adjectives turn up in letter after letter.
• Wrong facts about the employer. Outdated news, products the company no longer sells, offices that have closed.
• Overconfident legal and visa advice. Rules change and models can be out of date. Check official sources.
• Keyword stuffing. Asked to “optimise for ATS”, a tool may cram in every term from the ad, including ones you can’t support.
• Mass applying. Tools that fire off hundreds of applications send low-match CVs, can break job sites’ terms of use, and rarely raise your interview rate. Our look at how many applications it takes to get an interview explains why targeting matters more than volume.

1. You are the source. Give the tool your real CV, notes and results, and ask it to rearrange and rephrase, not to create.
2. Every claim must survive an interview. If you couldn’t talk about a line for two minutes, cut it.
3. Keep your structure and your voice. Use suggestions to edit your own draft rather than replacing it wholesale.
4. Verify facts yourself. Company details from the company’s own site; legal and visa rules from official sources.
5. Follow the rules you’re given. If an employer says no AI in an assessment or interview, don’t use it.

These are the same lines our guide to tailoring a CV without lying draws for any kind of editing. AI just makes it faster to cross them without noticing.

The pattern: AI handles volume and wording; judgement, facts and motivation stay with you.

About half an hour per application, with AI doing the mechanical parts and you making the decisions:

1. Triage (3 minutes): is the job open to you, in terms of location, visa and language? If not, stop here.
2. Match (5 minutes): compare your CV with the ad and decide whether you meet enough of the must-haves to apply.
3. Tailor (10 minutes): adjust your summary, the order of your skills and two or three bullets. Let AI suggest; you decide.
4. Letter (10 minutes): have AI draft from your notes, then rewrite the first and last lines yourself. Our cover letter examples show what a specific letter looks like.
5. Review (3 minutes): read everything out loud and check every number, name and date.
6. Apply and log it: the FindAJobAI Chrome extension fills repetitive form fields from your profile and never submits for you. Note the date so you know when to follow up.

Most problems come from open-ended prompts like “make my CV better”. Prompts that limit the tool to your own material produce safer, more useful output. Four to adapt:

1. Requirements check
"Here is a job ad and my CV. List the ad's requirements. For each one, quote the
line from my CV that supports it, or write 'no evidence'. Don't add anything
that isn't in my CV."

2. Bullet rewrite
"Rewrite this bullet so it starts with a verb and ends with the result. Keep every
fact and number exactly as given. If a number is missing, leave [blank] instead
of inventing one."

3. Cover letter draft
"Draft a 250-word cover letter for this job using only the achievements listed
below. Don't use the phrases 'I am writing to express' or 'passionate'. Mark
anything you're unsure about with [check]."

4. Interview practice
"Act as the hiring manager for this job. Ask me one behavioural question at a
time, wait for my answer, then tell me which part of the STAR structure was
weakest and what was missing."

Whatever you use, compare the output with your original line by line. Any number, tool or title that wasn’t in your input should be treated as invented until you’ve confirmed it.

Recruiters don’t need detection software to spot unedited AI text. The giveaways are familiar:

• Openers such as “I am writing to express my keen interest in…”.
• Stacks of adjectives: dynamic, passionate, results-driven, detail-oriented.
• Claims in the letter that don’t appear anywhere on the CV.
• Smooth sentences that could apply to any company, and wrong details about this one.
• Near-identical letters for different roles at the same employer.

Detection tools are unreliable, and the real test is simpler: is the application specific and true? A precise letter written with AI help reads fine. A generic one doesn’t, however it was written.

• Remove what a tool doesn’t need before you paste or upload: ID numbers, date of birth, full address, references’ contact details.
• Check what happens to your data: how long it’s kept, whether it’s used to train models, and whether you can delete it.
• Be wary of browser extensions that ask to read every website without a clear reason, and review their permissions.
• Watch for AI-assisted scams: fake recruiters with polished messages, interviews held only over chat, and offers that ask for payment or bank details early.

• It works from your documents. Output should be built from your CV and notes, not from a generic profile of your job title.
• It shows what changed. You should be able to see every edit next to your original before anything is saved or sent.
• You stay in control of sending. Be cautious with anything that applies on your behalf without your review.
• It has a clear data policy, including how to delete your account and files.
• Its claims are modest. No tool can guarantee interviews or “beat the ATS”; promises like that are a reason for doubt.

Practising with AI is one of its best uses. Have it generate likely questions from the ad, rehearse your answers out loud, and ask for feedback on structure and length. Our interview questions by job pages give you a starting list for your role.

For live interviews and assessments, follow the employer’s rules. Some explicitly ban AI tools during interviews or tests, and using them anyway can end your application. Either way, your stories have to come from your own experience: interviewers ask follow-up questions, and only real experience holds up under them.

Since we build an AI tool, here is what ours does and doesn’t do:

• Match scores. We compare your CV with each job using fixed rules plus an AI model, and show a 0–100 score. A score is an estimate to help you prioritise, not a verdict on you.
• Tailoring. Suggested changes keep your section order, your skills and their grouping, and your summary with small edits; bullets that don’t need to change stay exactly as you wrote them. It’s built not to add claims, and job titles are never added as skills.
• Autofill. The Chrome extension fills application forms from your profile. It never submits an application for you and never answers equal-opportunity or demographic questions.
• The limit. Models still make mistakes, so you review every change before you use it. That step isn’t optional, with our tool or anyone else’s.