New · AI

Tell me your experience and I'll write your CV

Speak or type however you like, in no particular order: where you've worked, what you did, where you studied. We turn that into a professional CV. We never make things up — anything missing, we'll ask you.

Your first voice generation is included: your story, the questions and a finished CV.

Optional. With your job, the AI uses the trade's vocabulary and asks what recruiters value.

Tell us a little more (a recording or a few lines) so we can start.

Used only to build your CV. You can edit everything afterwards.

01

What the AI does with what you tell it, step by step

The process has three phases, and understanding them is where the quality comes from. In the first, you supply the raw material: type it, paste an old CV, or dictate it out loud — usually the quickest route, and the most natural one for anyone who has spent years in a trade and never had to write their experience down.

In the second, that material becomes structure. The AI separates what you said into roles, dates, employers, duties, education, languages and skills, and flags the gaps: what is missing for the CV to be complete and competitive. If you said «I was at a city-centre restaurant for four years», the gaps are the venue name, the exact job title and the start and end months.

In the third, once those gaps are closed with a few short questions, the document gets written: headline, professional summary, achievements per role and section order. The draft is checked against the ATS criteria and, if it falls short of target, revised once more with the specific corrections needed. Then it opens in the editor, where you are in charge: every sentence is editable.

  • Any source: text, an old CV or dictated audio.
  • Extraction into structure, with gaps ranked by importance.
  • Short questions only for what is genuinely missing.
  • Drafting, ATS quality check and handover to the editor.

02

Why dictating beats filling in a form

Faced with an empty box labelled «work experience», almost everyone writes the poorest version of their job: two or three generic duties copied from what they assume is expected. Speaking does the opposite. Describing your shift surfaces the details that count: how many people you ran, how many orders went out, what went wrong and how you fixed it, what people asked you when there was a problem.

Those details are the raw material of an achievement. «Shift supervisor» says nothing; «shift supervisor with six staff and daily till reconciliation of around £2,000» places your level of responsibility in one sentence. The AI cannot invent those magnitudes — it can only organise and write the ones you provide.

One practical tip: dictate role by role, not all at once. Say the job title, the employer, the city, the months, what a normal day involved and what you were proud of. Do that and each role arrives with achievements already in place, with nothing to rewrite afterwards.

03

What AI does well, and what it must never do

It does three things well. Structuring: it turns a rambling account into sections with consistent dates and reverse-chronological order. Writing: it moves «helped in the warehouse» to a professional line with an action verb and a result. And tailoring: it rewrites the same history in the vocabulary of the role you are targeting, which is what shifts your match to the advert most.

There are two things it must never do. Invent experience you do not have: a fabricated job, employer, degree or certificate surfaces at interview or in reference checks, and the process ends there. And inflate figures: if you do not know how many customers you served, give an honest range or leave the number out.

One rule settles nearly every doubt: the AI writes, you answer. Anything left in the document must be something you can defend out loud for two minutes. If a sentence makes you hesitate, delete it before you send — it is never worth it.

  • Yes: structure, professional wording and tailoring to the role.
  • Yes: turning duties into achievements using your own figures.
  • No: job titles, employers, degrees or certificates that do not exist.
  • No: invented round percentages to look quantified.

04

How to review an AI-written CV in five minutes

Start with the hard facts: name, phone, email, city, dates and employer names. They are the one thing that allows no nuance and the first thing anyone checks. One wrong digit in your phone number turns a strong CV into an application nobody can contact.

Move to the headline and summary. The headline should state the role you want in the words used in adverts in your sector, not the internal title your old employer used. The summary has to sound like you: if you read a sentence you would not say out loud at interview, replace it with your own.

Then the achievements. Check one by one that the figure is real and that the verb describes what you did, not what your team did without you. Finally, check length and ordering: one page under ten years of experience, with your strongest material at the top, because a first read rarely gets past the opening third.

05

One history, one CV per advert

The most expensive mistake in a job search is sending the same PDF to thirty different adverts. Two postings with the same nominal title ask for different things, and the CV that wins is the one that speaks to what that employer actually wrote.

Tailoring does not mean rewriting your life: you adjust the headline, reorder achievements so the one matching the advert sits at the top, and align the vocabulary of your skills. With your history already structured, that is two minutes per application, not an afternoon.

Keep one well-built master version and generate each variant from it. That keeps everything consistent — same dates, same employers, same details — and avoids the classic problem of holding five different CVs and forgetting which one you sent where.

Have this to hand before you hit generate

  • Exact job title and employer for each role, as written in your contract.
  • Start and end month and year for every period.
  • One magnitude per role: people, volume, value or frequency.
  • Qualifications with their official name and awarding institution.
  • Languages at their true level, with a certificate if you hold one.
  • Licences and tickets the role requires (driving, forklift, food hygiene).
  • The role you are applying for, written as adverts word it.
  • A one-line explanation for every employment gap.

Common mistakes when using AI for your CV

Giving it four lines of information

With thin material, only a generic, empty CV can come out.

Dictate two minutes per role describing a normal working day.

Accepting the text unread

Any inaccuracy is yours to defend at interview.

Check hard facts, headline, summary and every figure before download.

Leaving filler language in

«Proactive and dynamic individual» separates nobody from anybody.

Swap it for a fact: what you did, with what, to what result.

One CV for everything

You lose match against each specific advert you apply to.

Adjust headline and achievement order per application from your master.

Translating word for word

Qualification and job titles do not map across markets.

Use the local equivalent and drop the photo and personal details.

Stretching to three pages

The important material sinks below where the first read stops.

Cut to one or two pages and lift your two best achievements up.

Questions about building a CV with AI