Best AI CV Builders UK 2026: 10 Tested by a Recruiter

Why most AI resume tools make your CV worse

In 12 years on the recruitment desk I’ve read somewhere north of 40,000 CVs. I can tell you the exact week ChatGPT changed the inbox. It was late November 2022. By February 2023, roughly one in three CVs that hit my desk opened with “results-driven professional with a proven track record”. By the summer the number was closer to one in two. The phrasing wasn’t new — recruiters had been wincing at it for a decade — but the volume was. AI didn’t invent CV cliché. It mass-produced it.

The problem with most AI CV tools isn’t that they write badly. It’s that they write the same. Feed ten different candidates the same prompt and you get ten near-identical drafts. Same opening line, same verb stack (“spearheaded, drove, delivered, optimised”), same three-bullet structure under each role. The CV stops doing the one job it has, which is to make a hiring manager pause for two extra seconds and read the next line.

The good news: AI is genuinely useful for CV writing — for brainstorming bullets, for tightening sentences, for catching the bit where you’ve written “managed a team” five times in one document. The bad news: the way most candidates use it produces a CV that’s slightly worse than the one they’d have written on a Sunday afternoon with a cup of tea. This guide is the recruiter’s-eye view of which tools help, which hurt, and the editing habits that turn an AI draft into a CV that actually gets a callback.

If you remember nothing else from this page: AI drafts. You edit. Not the other way around. Every section below comes back to that rule.

What recruiters actually do in the first 8 seconds

You’ve probably read the 8-second statistic before. It comes from eye-tracking research and it’s directionally true, though the reality is messier. When a CV opens on my screen, here’s what actually happens — and I’ve timed myself doing this hundreds of times.

Want to see what your shortlist probability actually is? Paste your CV + a UK job description into the free Will I Get Shortlisted? tool — it returns a recruiter-calibrated probability percent, the verdict from a 12-year UK recruiter, and three specific fixes that lift you above the cut. No signup, no AI hallucination — pure rules engine, 100% client-side.

Seconds 1–2: the top third. I look at three things, in this order: your current job title, the company name underneath it, and the dates. That’s it. If you’re applying for a Senior Product Manager role and your current title is “Product Lead” at a company I recognise, you’ve earned the next six seconds. If your current title is “Founder” of something I’ve never heard of, or “Career Break” with no context, my brain has already started preparing a polite rejection.

Seconds 3–4: the headline or summary. I don’t read it word for word. I scan it for two or three nouns that match the job spec — domain (fintech, SaaS, healthcare), function (growth, infrastructure, payments), seniority signal (led a team of, P&L of). If those nouns are missing or buried under three lines of “passionate, dynamic professional”, I skip the summary entirely and drop to your second job.

Seconds 5–6: bullet density and verbs. I’m not reading bullets yet. I’m scanning the shape of them. Are they short and metric-led? Or are they long, narrative, and abstract? A CV where every bullet starts with “Responsible for” tells me you’ve described your job rather than what you achieved in it. A CV where every bullet starts with “Led, drove, owned” tells me the opposite — sometimes accurately, sometimes because ChatGPT was asked to write it. Run any bullet through the UK CV Bullet Quality Scorer before submitting — it scores 0-100 on metric strength, action verb, and length.

Seconds 7–8: the second and third roles. This is where I check whether the career story makes sense. Two-year tenure at the current job is fine. Four jobs in three years across unrelated industries needs an explanation, and your CV should pre-empt it.

What I’m not doing in those eight seconds: reading your education section, looking at your hobbies, admiring your design template, or counting your buzzwords. The CV is a triage document. The eight-second pass decides whether you go in the “yes” pile, the “maybe” pile, or the “no” pile. About 70% of CVs land in “no” on the first pass. Of the remaining 30%, maybe half make it to the proper read.

This matters for AI tools because most of them optimise for the wrong thing. They optimise for keyword density, ATS parsing scores, design polish — none of which I see in eight seconds. What I see is your title, your dates, and the rhythm of your bullets. If you’re using an AI tool, ask yourself: does this thing improve the top third of my CV? Most of them don’t. I’ve broken the failure modes down in why AI-written CVs get caught, but the headline is that AI tends to bloat the parts of the CV I scan and leave the parts I scrutinise (your achievements) untouched.

The second eight-second pass — when I come back to a “maybe” pile CV later in the day — is where bullets get read properly. That’s where AI either earns its keep or sinks you. We’ll come back to that.

The 6 ways AI ruins a CV — and how to use it without getting caught

Here’s the framework I use when I’m coaching candidates I’ve placed. Six failure patterns. If your AI-written CV has even three of these, a hiring manager with any pattern recognition will spot the help.

1. The verb stack collapse. Every bullet starts with the same five verbs: led, drove, spearheaded, leveraged, optimised. ChatGPT is statistically biased towards these because they appear in nearly every CV in its training data. The fix: print your CV, highlight the first word of every bullet, and count repeats. If “led” appears more than twice on a one-page CV, you have the verb stack collapse. Replace with verbs that match the actual work: “rebuilt”, “negotiated”, “shipped”, “diagnosed”, “killed” (as in “killed the £40k vendor contract that wasn’t paying back”). Specific verbs beat strong-sounding verbs every time.

2. The metric-shaped hole. AI loves the shape of a metric without the substance. “Improved customer satisfaction by leveraging cross-functional initiatives” — sounds quantified, isn’t. Real metric: “cut customer onboarding time from 12 days to 3, which reduced churn in the first 90 days by 18%”. The first version is what AI writes by default. The second is what you have to add yourself, because the AI doesn’t know your numbers. If your CV has metric-shaped sentences with no numbers in them, you’ve got the metric-shaped hole.

3. Buzzword inflation. Six years ago a candidate would have written “managed the team”. After two rounds of ChatGPT, that becomes “orchestrated cross-functional team dynamics to drive collaborative excellence”. I have a running list of phrases I delete on sight — synergy, holistic, robust, results-driven, dynamic professional, passionate about, thought leader. If your CV has more than two of these per page, AI bloated it.

4. The achievement-without-context bullet. “Delivered 30% revenue growth.” Cool. From what baseline? Over what period? Was that at a £2m business or a £200m business? Were you the one who delivered it, or part of a team of forty? AI writes these bullets because the prompt didn’t include the context, and the model fills the gap with something that sounds impressive. Recruiters discount these bullets to nearly zero. Always include scope and timeframe: “Grew the EMEA pipeline from £4m to £5.2m (30%) in 14 months as the only enterprise rep covering DACH.”

5. The same-shape paragraph in the summary. AI summaries follow a template: one sentence on years of experience, one on industry breadth, one on a soft skill, one on a desired outcome. Any recruiter who reads CVs all day can spot this rhythm in two seconds. The fix is structural — write the summary as three short, punchy lines that say what you do, what makes you different, and what you want next. No template, no flow, just signal.

6. The tense and pronoun drift. Watch what happens when you paste three roles into ChatGPT and ask it to write bullets: tenses wander between past and present, occasionally drifting into “I” and “we”. A human writing their own CV doesn’t do this — they’re consistent. The drift is a tell. After any AI pass, do a single read-through where the only thing you check is tense (past for past roles, present for current) and pronoun (none — CVs don’t use “I”).

The way to use AI without falling into these traps is to treat it as a first-draft accelerant, not a writer. Get it to brainstorm twelve bullets per role. You’ll keep four. Get it to suggest three different summary openings. You’ll write your own using the best phrase from one of them. Get it to spot weaknesses (“which of these bullets is the weakest and why?”). It’s surprisingly good at that diagnostic work, and bad at the generative work people use it for. There’s more on this distinction in can recruiters tell if you used AI — short answer: yes, but only when you let the AI keep the pen.

The UK CV format that survives the ATS in 2026

If you’re applying in the UK in 2026, your CV should be a Word .docx file (not PDF, not pages, not a Canva export), two pages maximum, single column, in Calibri or Arial 10–11pt. I’ll defend each of those choices in turn.

File format. Most modern ATS platforms (Workday, Greenhouse, SmartRecruiters, Lever, iCIMS) parse .docx more reliably than PDF. PDFs can parse, but they fail more often, especially when the original was exported from a design tool. The exception is Workday, which strips formatting from anything you upload anyway. Default to .docx. Name the file Surname-Forename-CV.docx — not “CV final v3.docx” or “Resume-2026-updated.docx”. Recruiters search their downloads folder by candidate name; help us find you.

Length. Two pages is the UK standard. One page if you have under five years’ experience. Three pages only if you’re in academia, medicine, or you’ve genuinely had a 25-year career with relevant detail at every stage. The American one-page rule doesn’t apply here — UK recruiters expect more context, and a one-page CV from a senior candidate often reads as undersold rather than concise. There’s a fuller breakdown in the UK CV format 2026 guide.

Structure. The order I recommend, top to bottom: contact details (name, phone, email, LinkedIn URL, city — no full address), professional summary (3 lines max), experience (most recent first, reverse chronological), education, key skills, optional sections (languages, certifications, volunteering). No photo. No date of birth. No marital status. No “References available on request” — it’s wasted space and the assumption is they’re available.

The single-column rule. Two-column CVs look elegant in a design tool and parse appallingly in the ATS. The parser reads top to bottom, then left to right, so a two-column CV often comes out with your education sandwiched into the middle of your most recent role, or your phone number landing in the middle of a bullet point. Use a single-column layout. If you want visual hierarchy, use bold headings and white space, not columns.

Fonts. Calibri 11, Arial 10, or Helvetica 10 for body. Headings can go up to 12 or 13 bold. Avoid: Times New Roman (reads as dated in 2026), Garamond (parses inconsistently), Comic Sans (don’t), and any custom font that won’t be installed on the recruiter’s machine — the file falls back to Calibri anyway, and your spacing breaks.

Sections to include. Professional summary, work experience, education, key skills. Optional: certifications (only if relevant), languages (only if proficient enough to use professionally), volunteering (only if it adds a skill or industry signal the rest of the CV doesn’t), publications (academic or technical roles only).

Sections to delete. Hobbies, unless they’re load-bearing (a marathon runner applying for an endurance-sports brand, fine). Personal statement that duplicates the professional summary. “Career objective” — the objective is the job you’re applying for. Photo. Date of birth. Nationality (in the UK, employers don’t ask, and including it triggers right-to-work questions you don’t want).

Keyword placement. The ATS scores keyword matches with weighting that favours the top of the CV. A keyword in your professional summary scores higher than the same keyword in your fourth bullet of role three. This is why generic AI tools fail — they distribute keywords evenly. You want them concentrated at the top, in the summary and the most recent role, fading naturally as you move down the CV. There’s a method to this — see the keyword-match section below and the free CV keyword match score tool.

The recruiter version vs the ATS version. A CV that looks beautiful to a human can be invisible to a parser. A CV that scores 95% on an ATS keyword match can read as desperate and stuffed to a human. You’re optimising for both audiences, and they want different things. The single-column .docx with concentrated-but-natural keywords at the top is the format that survives both. Anything more ambitious is a gamble.

Tool-by-tool: which AI resume builders actually work

I’ve tested every major tool on this list against the same brief — a UK-based senior product manager, fintech background, applying for a Head of Product role. Here are the verdicts. For direct head-to-head matchups across the whole site, see the JobLabs comparisons hub.

Teal. The strongest all-rounder for active job seekers. Tracks applications, generates per-job tailored versions of your CV, has a usable AI bullet generator. The bullet AI is better than ChatGPT out of the box because it’s trained on a narrower set of CV-specific patterns. Verdict: worth the subscription if you’re applying to more than ten jobs a month. Free tier is genuinely usable for the first few applications. Compared head-to-head in Teal vs Rezi and ChatGPT vs Teal.

Rezi. ATS-first builder with strong template parsing. The interface forces you into clean structure, which prevents the formatting disasters I see on Canva-built CVs. The AI content generator is weaker than Teal’s — bullets come out generic and need heavy editing. Verdict: best for candidates who need to fix structural problems, less useful if your CV is already well-formatted. Strong on the ATS side, weaker on the writing side.

Jobscan. Not a builder — a keyword-match scorer. You paste your CV and the job description, and it tells you which keywords are missing. Best used as a checker, not a writer. The score is directionally useful but not gospel — I’ve seen Jobscan reject CVs that recruiters loved and approve CVs that were keyword-stuffed nonsense. Use it as one signal, not the final word. See Jobscan vs Resume Worded for the head-to-head.

Resume.io. Polished templates, weak AI writing. The templates are tempting because they look professional, but several of them are two-column designs that destroy ATS parsing. If you use Resume.io, pick the simplest single-column template, ignore the AI suggestions, and treat it as a glorified Word document. Verdict: visually appealing, structurally risky.

Resume Worded. Strongest in the audit category — it scores your CV against ATS and recruiter criteria with line-by-line feedback. Less useful as a builder; the suggestions tend toward generic. Use the score as a diagnostic, ignore the rewrites unless you’re stuck. Best paired with another tool that does the actual writing.

Kickresume. Template-heavy, AI-light. The AI suggestions are weak and often produce buzzword-loaded bullets. Templates are decent but skew design-heavy in ways that hurt ATS parsing. Verdict: skip unless you specifically want the design polish and you’re confident your industry doesn’t go through an aggressive ATS (creative industries, mostly).

ChatGPT (and Claude). The free, flexible option. With the right prompts, ChatGPT writes better bullets than most paid tools — but the prompts have to be specific. “Write me a CV bullet about leading a product launch” produces garbage. “Rewrite this bullet to lead with a quantified outcome, use a specific verb, and stay under 22 words: [paste bullet]” produces something usable. The full prompt library is in ChatGPT prompts for resumes. Verdict: the strongest tool in skilled hands, the worst in unskilled ones.

Grammarly. Not a CV builder, but the best polish layer. Run your final draft through Grammarly’s clarity and conciseness checks and you’ll catch the kind of bloat that survives even careful editing. Use it last, not first.

The honest summary. If you can only pay for one tool, Teal. If you can pay for two, Teal plus Jobscan as a keyword checker. If you’re broke, ChatGPT plus a free Grammarly account will get you 80% of the way for £0. The expensive premium tools (Kickresume, Resume.io paid tiers) don’t earn their keep. Pick from the full builder review if you want the longer breakdown.

What no tool does well: write your career story. Every tool above is a writing accelerant, not a thinking accelerant. The thinking — what made you good at your last job, what you actually want next, which two achievements are worth fighting for column inches — is yours. Tools that promise to do the thinking for you are the ones that produce the templated output recruiters reject.

The keyword-match trap (and how to escape it)

There is a particular kind of CV death that comes from over-optimising for keyword match. I see it constantly: a candidate scores 92% on Jobscan, 95% on Resume Worded, and zero callbacks. The CV reads like keyword soup. They’ve stuffed every term from the job spec into bullets that no longer describe real work.

Here’s the truth about ATS keyword matching in 2026. It still matters, but it’s not the gating factor it was five years ago. Most modern ATS platforms now use semantic matching — they understand that “managed a P&L” and “owned the budget” mean roughly the same thing. The game is no longer “include the exact words from the job spec verbatim”. It’s “include enough domain-relevant language that the parser knows you’re in the right ballpark, then write bullets a human will actually read”.

The mistake candidates make under AI guidance is treating the ATS like a 2018 ATS — keyword-density-driven, exact-match-required, formatting-fragile. That ATS is mostly gone. What replaced it is a system that scores you on relevance, then dumps a shortlist of 30–50 candidates onto a recruiter’s desk. Your CV has to clear the parser, which is the easy part. Then it has to clear the recruiter, which is the hard part. The recruiter is reading for substance. Keyword stuffing — what I’ve seen called keyword stuffing in the trade — actively hurts at the recruiter stage.

Here’s how to do keyword matching properly in 2026.

Step 1: extract the real keywords. Paste the job description into the free CV keyword match score tool. Look at the keywords it surfaces. Ignore the soft skills (communication, leadership, teamwork) — those are noise. Focus on the hard ones: tools, methodologies, domains, certifications, specific outcomes the job mentions.

Step 2: cluster them. Group the keywords into three buckets. Bucket one: terms you’ve genuinely used in past roles, with evidence. Bucket two: terms that describe work you’ve done, but using different language in your current CV. Bucket three: terms that describe work you haven’t actually done.

Step 3: rewrite, don’t insert. For bucket-one terms, make sure they appear in the natural flow of your CV. For bucket-two terms, rewrite the relevant bullets to use the job-spec language where it’s accurate (this is where most of the work is). For bucket-three terms, leave them out entirely — putting work on your CV that you haven’t done is a faster route to rejection at interview than to rejection by an ATS.

Step 4: re-score and stop. Run the match score once. Aim for 65–75%. Anything above 80% on a real CV is suspicious — it usually means you’ve stuffed. Anything below 50% means you haven’t done the rewriting work. The sweet spot is the boring middle.

The sequence matters because most candidates do the inverse. They run the score first, see they’re at 30%, and panic-stuff. The CV ends up at 88% match and 0% recruiter response. Score is a diagnostic, not a target. There’s a longer treatment of the parser side in how the ATS really works, and a free tool above that gives you the directional read without the urge to over-optimise.

One more thing about keywords: location. The same keyword counts for more in your professional summary than in a bullet on page two. The ATS weights position. So if a job spec mentions “subscription pricing” and you’ve worked on that, the phrase needs to appear in your summary and in the most recent role’s bullets — not buried on page two in a job from 2017.

The hybrid workflow recruiters can’t spot

Here is the workflow I give every candidate I coach. It uses ChatGPT (or Claude — pick one) for the heavy lifting and produces a CV that doesn’t read as AI-generated. It takes about three hours for a full rewrite, ninety minutes for tailoring an existing CV to a new role.

Stage 1: brain dump (45 minutes, no AI). Open a blank document. For each role on your CV, write down everything you did, in plain English, without trying to make it sound impressive. Include numbers wherever you remember them — even rough ones. “Ran a team of about 12, hit our annual target two years out of three, the third year we missed because the product launch slipped.” This is your raw material. It’s the bit AI cannot do for you, because the model doesn’t know your numbers. If you skip this stage, every later stage produces generic output.

Stage 2: AI bullet generation (30 minutes). Feed each role’s brain dump into ChatGPT with this prompt: “Below is a brain dump of what I did in my role as [title] at [company]. Generate 8 CV bullets, each starting with a strong specific verb, each containing at least one quantified outcome, each under 22 words. Avoid the verbs led, drove, spearheaded, leveraged. Avoid the words synergy, holistic, robust, results-driven.” The constraints matter — without them you get the verb stack collapse from earlier. This produces 8 bullets per role; you’ll keep 3 or 4.

Stage 3: human edit (60 minutes). This is the load-bearing stage. For each bullet ChatGPT produced, ask three questions: is this true, is this specific, is this in my voice? Most of the bullets will fail one of those tests. Rewrite them in your own words. Add the context the AI missed. Cut the ones that are too generic to save. By the end of this stage you should have 3–4 bullets per role that you’d be willing to defend in an interview.

Stage 4: summary and headline (20 minutes, mostly human). The summary is the bit AI is worst at. Write three versions yourself first. Then ask ChatGPT to “tighten this without changing the substance, removing any cliché phrases or buzzwords”. You’re using the AI as an editor, not a writer. The output should be 90% your words and 10% AI tightening.

Stage 5: keyword pass (15 minutes). Run your draft through the keyword match tool. Identify three or four genuine keywords that aren’t in the CV. Rewrite specific bullets to include them where accurate. Don’t insert them artificially.

Stage 6: Grammarly polish (10 minutes). Run the whole thing through Grammarly. Accept clarity and conciseness suggestions, ignore tone suggestions (Grammarly defaults to making things sound friendlier; CVs should sound competent, not warm).

Stage 7: read aloud (5 minutes). This is the test that catches everything else. Read your CV aloud, slowly. Anywhere your mouth trips over a phrase, anywhere a sentence sounds like something you’d never actually say in conversation, edit. AI text trips you up here because it has the wrong rhythm. Human text doesn’t.

The whole process is about three hours for a from-scratch rewrite. The reason it works is that AI is doing the parts AI is good at (generating options, tightening prose, running checks) and you’re doing the parts AI is bad at (telling the truth, picking the strongest version, sounding like yourself). The bullets that come out the other side are stronger than what you’d write alone, and they don’t pattern-match to AI output.

What this beats: the candidate who pastes a job spec into ChatGPT and asks for “a CV”. What it doesn’t beat: a candidate who uses the same workflow but actually has the experience the job needs. AI can polish your CV. It cannot make you qualified. There’s a fuller version of the tailoring side in how to tailor your resume to a job description.

Career changers: the special case

If you’re changing careers, the rules above mostly apply, with three exceptions that matter.

The summary does more work. A normal CV summary signals fit through job titles and industry. A career-changer’s CV doesn’t have that on its side — your job titles and industry are about to change. The summary has to do the work of explaining the move. Three lines: who you’ve been, what you’re moving towards, the bridge between them. Example: “Six years in management consulting, three of those building digital transformation programmes for retailers. Now moving full-time into product management in retail tech, where I’ve spent the last 18 months as the business stakeholder on a £4m platform rebuild.” The bridge is the third line — the bit that says “I’m not random, I’ve been adjacent to the new role for a while”.

Transferable skills are real, but you have to translate them. A consultant moving into product can claim “stakeholder management, cross-functional delivery, prioritisation under constraint” — those are transferable. They have to be rewritten in product-team language, though, not consulting language. “Drove client-side delivery against a fixed scope” is consulting. “Shipped against quarterly OKRs with cross-functional team alignment” is product. Same skill, different vocabulary. AI is genuinely useful here — paste the consulting bullet, ask ChatGPT to rewrite it in product-management language, edit. Just don’t trust the first output.

Lead with proximity, not aspiration. The strongest career-change CVs lead with the project, side hustle, course, or volunteer role that’s closest to the new field — not with the existing job title. If you’re moving from teaching into UX research, and you’ve done a UX bootcamp and three pro bono projects, those projects belong in the top third of your CV, above your teaching role. The chronology purists will hate this; they’re wrong. Recruiters don’t care about chronology, they care about relevance. The full treatment is in AI resume for career changers.

The reason career changers get bad AI output is that the prompts they use don’t acknowledge the change. “Write me a CV bullet for my teaching job” gives you a teacher’s bullet. “Write me a CV bullet about my teaching job that emphasises stakeholder management, communication under pressure, and rapid iteration on feedback — three skills relevant to UX research” gives you something useful. The prompt has to do the translation work for the model to do it well.

Career changers also have a higher bar to clear at the screening interview, because every recruiter on the other end is trying to assess whether you actually understand the new field or you’re aspirationally applying. Your CV has to pre-empt that by demonstrating proximity rather than claiming it.

The 30-second pre-flight checklist before you submit

Before you click submit on any job application, run through this list. It takes thirty seconds and catches roughly 80% of the mistakes I see in the inbox.

1. File name is Surname-Forename-CV.docx, not “CV final v3” or “Resume - updated”.
2. Top third has your current title, current company, and a 3-line summary that names your function and domain in the first ten words.
3. First word of every bullet is a verb, not “Responsible for” or “Worked on”.
4. No bullet starts with the same verb as the bullet above it. Check the first words top to bottom — if any verb repeats more than twice on the page, swap.
5. Every bullet has a number, a name, or a specific noun. A bullet without one of those three is decoration, not evidence.
6. No bullet runs to three lines. If it does, it’s two ideas, split it.
7. Buzzword check. Search the document for: synergy, holistic, robust, results-driven, dynamic, passionate about, thought leader, leveraged. Delete or rewrite each one.
8. Tense check. Past roles in past tense, current role in present tense. No mixed tenses inside a single role.
9. No “I” or “we” anywhere on the CV.
10. Page count. Two pages, single column. Not 1.5, not 2.5. Tighten or expand to a clean two.
11. Contact details include phone, email, LinkedIn URL, city. No full address, no date of birth, no photo.
12. Job-spec keywords. Open the job spec in another tab. Skim it. Check that the three or four most prominent terms appear naturally somewhere on your CV — ideally in the summary and most recent role.
13. Read the first six lines aloud. If anything sounds like marketing copy, rewrite.
14. Spellcheck on, but don’t trust it. It misses “manger” for “manager”, “pubic” for “public”, and other classics. Eyeball the headings.
15. Save and close. Reopen the file. Confirm it actually opens cleanly in Word, no formatting drift.

Most rejections aren’t dramatic. They’re a tired recruiter at 4pm on a Thursday seeing one of the items above and deciding it’s a “no”. The pre-flight is cheap, the cost of skipping it is invisible — you’ll just never know which “no” you caused yourself.

If you want a deeper version of this checklist, the pre-flight in the UK CV format guide goes further on formatting specifics.

What to do if your CV is getting zero callbacks

Here is the diagnostic flow I use when a candidate tells me they’ve sent out 50 applications and heard nothing. It’s not a single problem. It’s usually one of four, and the fix depends on which.

Diagnosis 1: ATS rejection. You’re not even reaching a human. Symptom: zero responses, including auto-rejections that come back within 48 hours. Cause: usually a structural CV problem (two-column layout, image-based PDF, weird font) or a keyword mismatch so severe the ATS doesn’t shortlist you. Fix: rebuild the CV in a single-column .docx, run it through the free keyword match scorer against three real job specs, and check you’re hitting 60%+ on each. If you’re hitting under 40%, the CV is being filtered out before any human sees it.

Diagnosis 2: human rejection at first scan. You’re getting through the ATS but not getting calls. Symptom: most applications go silent, occasional auto-rejection comes back days later (a human pulled it from the shortlist). Cause: the CV passes the parser but the recruiter’s eight-second scan turns up nothing. Fix: rebuild the top third. Title + company + summary needs to be doing the work. If your current title is ambiguous (“Lead”, “Head of”) and your company is unknown, the summary has to compensate by naming your domain and scale of impact in the first line.

Diagnosis 3: relevance mismatch. You’re applying for jobs you’re not quite qualified for. Symptom: callbacks for some roles, silence for others. Cause: you’re aspirationally applying to roles a tier above your evidence. The CV is fine; the targeting isn’t. Fix: cut the top 20% of jobs you’re applying to, the ones where you’re hoping the CV stretches. Apply to the ones where the fit is obvious. Once you have callbacks landing, you can stretch from a stronger position. There’s a longer breakdown in why am I not getting hired covering the eleven specific rejection reasons I see most often.

Diagnosis 4: market reality. Sometimes the CV is fine, the targeting is fine, and the market is just slow. Symptom: applications go silent across the board, peers report similar experiences, the role you’re hunting has thinned out on the job boards. Cause: hiring freeze, sector contraction, end-of-quarter budget pause. Fix: there isn’t one for the CV. The fix is patience plus a reduced application target, plus more direct outreach. CV optimisation cannot beat zero open roles.

The mistake candidates make is to assume diagnosis 1 (ATS) without checking the others, rebuild the CV three times, and end up with a worse document because they were solving the wrong problem. Run the diagnostics in order. ATS first (cheap to check), then human scan (rebuild the top third), then targeting (cut the stretch applications), then market (accept it).

If you’re stuck, the single most useful diagnostic is to send your CV to a recruiter in your industry and ask “would you call this candidate?” Not “is the CV good?” — recruiters will diplomatically say yes. “Would you call them?” forces a real answer. If you can’t reach a recruiter, paste the CV and the job spec into ChatGPT and ask “what would make you not call this candidate?”. The output is biased and AI-shaped, but it surfaces the obvious gaps. Tools like the keyword match scorer are useful as a second opinion, not a primary diagnosis.

A note on the wider job hunt — there’s a separate piece on the UK take-home pay calculator for the offer-stage maths, and the career-change route covers the tooling beyond the CV stage. The CV is the front door. Once you’re through it, the rest of the search is its own discipline.

The single rule that runs through this whole guide: AI is a tool, not a writer. Used as a tool — for brainstorming, tightening, scoring, checking — it makes your CV measurably better. Used as a writer, it makes your CV measurably worse. The candidates I’ve placed over the last three years who used AI well treated it like a junior researcher: useful, fast, requires supervision, never gets the final word. That’s the model. Everything above is detail.

Related reading

• UK CV format 2026 — the 7-section UK CV that survives the 8-second skim and parses cleanly through every ATS.
• CV examples by role (30 roles) — recruiter-approved bullet patterns and full CV structures for product manager, software engineer, designer, accountant, and 26 other UK roles.
• Why am I not getting hired? — 11 specific recruiter rejection reasons before you blame the CV.
• Can recruiters tell if you used AI? — the 8 dead giveaways and how to use AI without getting caught.
• AI tools for every phase of the job search — the full tool stack beyond just the CV builder, mapped to each stage of the search.

Where to go after the CV

The CV is one document in the application package. Once it’s working:

• Cover letter that survives the half-page skim — the cover letter pillar. Five opening patterns, the middle paragraph that actually moves a hiring decision, and how to write it without sounding like AI.
• Interview prep without sounding rehearsed — the interview pillar. STAR framework, the 4-stage UK process, AI-prep workflow, and the recruiter pro tips that turn a strong CV into an offer.
• LinkedIn profile that recruiters actually click — the LinkedIn pillar. Headline formula, About section structure, and the 5-section ordering that lets recruiters find you instead of waiting for you to apply.
• Career change playbook — the career-change pillar. How to lead with proximity not aspiration, the transferable-skills exercise, and the salary-step-back conversation when switching sectors.

Free tools to pair with the CV

The CV is one document but the application package usually needs supporting moves. These free recruiter-built tools are the ones I’d send candidates alongside any CV review:

• CV Keyword Match Score — score how well your CV matches the JD before you apply. ATS-stage filtering is brutal; this is the cheapest pre-screen you can run.
• Job Description Analyzer — decode the JD before tailoring the CV. Spot the hidden skills, the red flags, and the must-haves vs nice-to-haves.
• UK net pay calculator — what the gross figure on the JD actually puts in your pocket after tax, NI, and pension.
• UK Pay Rise Calculator — three recruiter-calibrated negotiation bands once the offer comes in.

UK reference guides for the wider context

• UK CV Format 2026 — the structural reference: 7-section CV format, ATS-safe layout, length conventions, and what each section does.
• UK Salary Guide 2026 — UK ranges by role + city, so the salary section of your application package is anchored in real numbers.
• UK Cover Letter Guide 2026 — the cover letter that gets sent with every CV.