AI Cover Letter Mistakes UK 2026: 8 Patterns Recruiters Bin Fast

I read between 30 and 50 cover letters a day. In 2026 maybe 60% of them are written by ChatGPT, Gemini, or whatever in-house AI tool the candidate has open. I can spot most of them inside eight seconds, and so can every other UK recruiter I know.

Here’s the thing: AI cover letters aren’t bad because they’re AI. They’re bad because the candidate didn’t edit them. A raw ChatGPT draft has a fingerprint, and once you’ve seen it 400 times you can’t unsee it. Below are the eight patterns I bin instantly, and the rewrite that gets the callback.

1. The “I am writing to express” opener

Nine in ten AI cover letters start with some flavour of:

“I am writing to express my keen interest in the Marketing Executive position at Acme Ltd.”

This is the line that triggers my eye-roll reflex. It tells me nothing. It says you found a job ad, you applied, and you have the linguistic register of a 1998 form letter. I read this opener probably 200 times a week.

The fix: open with one specific sentence that proves you read more than the job title.

“I switched my agency’s reporting to GA4 last year, so when I saw Acme is hiring a Marketing Executive specifically to migrate analytics platforms, I stopped scrolling.”

That’s a hook. That’s a reason for me to read paragraph two. The second version takes 30 seconds longer to write than the AI default, and it doubles your interview-call rate. I’ve tested this with my own candidates.

2. Buzzword soup

ChatGPT’s default register is corporate-LinkedIn — the same hollow voice you’ll spot on half the profiles in your feed. Ask it to write a cover letter and you’ll get “results-driven”, “passionate”, “dynamic”, “spearheaded”, “leveraged”, “synergy”, “cross-functional”, “holistic” — sometimes three of them in one sentence.

I don’t read past the second buzzword. Neither does any recruiter I trained.

The fix: ban every word from the list above. If your letter still works without them, you didn’t need them. If it doesn’t, your letter was buzzword scaffolding holding up nothing.

Replace “I leveraged my cross-functional skills to spearhead a results-driven campaign” with “I ran the Q3 campaign that brought in 1,200 leads at £14 each, 40% under our target CPL.”

One is a sentence. The other is a confession that you have nothing to say.

3. Generic company praise

Every AI letter has a paragraph that sounds like this:

“Acme Ltd’s commitment to innovation and excellence in the technology sector aligns perfectly with my values.”

This line works for Acme, BAE Systems, Greggs, and the local plumbing firm. That’s the problem. AI doesn’t know anything specific about the company you’re applying to, so it writes praise that fits any company. Recruiters spot this because we’ve read the same sentence about our own client and three of their competitors in the same week.

The fix: name a product, person, blog post, or news event from the company. Two sentences:

“I read your CTO’s post on the move from monorepo to microservices in March. The bit about gradual extraction over big-bang rewrites matched what I’ve been doing at my current shop, and I want to work somewhere that thinks that way.”

Now I know you actually looked at the company. That puts you ahead of 80% of applicants.

4. Fabricated or vague metrics

AI hallucinates numbers. I’ve seen letters claim “I increased team productivity by 47%” with no method, no baseline, no timeframe, and no plausible source for the figure. ChatGPT made it up. So did the candidate when they didn’t catch it.

The other version is the metric that means nothing: “I improved efficiency,” or “I drove significant growth.” What’s significant? Drove how? Efficiency in what?

The fix: real numbers, with context.

“In my last 18 months I cut our customer onboarding from 11 days to 4. The bottleneck was a manual contract step — I built a DocuSign template that auto-fills from Salesforce.”

Specific number. Specific timeframe. Specific method. A recruiter or hiring manager can ask you about it in interview, which is exactly what you want, because it proves you actually did it — and the recruiter-side interview prep that turns claims into offers builds on exactly the same evidence-first habit.

5. Em-dash plague

ChatGPT loves the em-dash. So do I, used sparingly. The AI default uses three or four per paragraph, often where a comma or full stop would do.

“My experience in marketing — particularly in digital channels — has prepared me — through both formal training and hands-on work — for this role.”

That sentence has four em-dashes and zero content. Real writers use one, maybe two, per page.

The fix: find every em-dash in your letter. Keep the best one. Replace the rest with full stops or commas. Your letter will read 30% cleaner immediately.

6. The list-of-three reflex

AI writes in threes. Always threes. “Reliable, hardworking, and team-oriented.” “Strategy, execution, and results.” “My passion, dedication, and expertise.”

Once you notice this you can’t stop noticing it. I see four or five tricolons in every AI letter, often back-to-back. It’s because the model has been trained on motivational LinkedIn posts where the rule of three rules everything.

The fix: count your “X, Y, and Z” structures. Cut at least half. Replace one with a single specific noun. “My experience in B2B SaaS pricing models” beats “my passion, dedication, and expertise” every time, because one tells me what you actually do.

7. The “perfectly aligned values” cliche

“Your company’s values align perfectly with my own.”

I’ve read this sentence so many times I’ve genuinely thought about printing it on a mug. It says nothing. It can’t say anything, because the AI doesn’t know either of your values. It’s a placeholder pretending to be a sentence.

The fix: if you genuinely care about something the company does, say it concretely. “You’re the only firm in the UK pensions space publishing your fee structure on the homepage. That’s why I want to work here, not at one of your competitors.” That’s an aligned value, expressed with evidence.

If you can’t write that sentence honestly, delete the whole “values” paragraph. A short letter without a fake values paragraph beats a long letter with one. See my guide on cover letter length for why 250-350 words usually wins.

8. The closing-line cliche

The AI default closer is one of three flavours:

“I would welcome the opportunity to discuss how my skills can contribute to your team.” “I look forward to the possibility of contributing to your continued success.” “Thank you for considering my application. I am eager to hear from you.”

I see all three multiple times a day. They’re the conversational equivalent of a polite cough. They give the reader nothing to act on.

The fix: close with a specific ask or a specific availability.

“I’m in London Tuesday and Thursday next week if a 20-minute call works. Otherwise happy to come to your Shoreditch office any morning before 11.”

That tells me you’re decisive, you’ve thought about logistics, and you actually want this job. It also makes booking the interview a one-line reply for me, which is exactly the friction you want to remove. My piece on cover letter opening lines covers the same logic for the top of the letter.

The 8-second test you can run yourself

Print your letter. Read only the first sentence and the last sentence. If both could appear, word-for-word, in 100 other people’s cover letters, you’ve written an AI letter regardless of who typed it.

Now read paragraph two. Find the most specific noun in it. If the most specific thing is “team” or “company” or “role”, you have a buzzword problem. If the most specific thing is a product name, a person, a number with units, or a date, you’re probably fine.

This is the same test I run when I’m screening, except I do it in eight seconds instead of two minutes.

Why this matters more in 2026 than it did in 2023

Two things have shifted. First, the volume of AI cover letters has tripled, so the recruiter pattern-recognition is sharper. We’ve all calibrated to the AI rhythm. Second, the candidates who edit their AI drafts properly now stand out more than they did three years ago, because the unedited ones drag down the average.

If you spend 20 minutes turning a ChatGPT draft into a human letter, you’re competing against people who spent 90 seconds. The maths is on your side. For the full guide to using AI without these tells, see how to write a cover letter with AI.

For more on the broader cover letter craft — formatting, tone, structure — start at the cover letter pillar and work outward.

Bottom line: AI cover letters fail because the candidate skipped the editing. The eight patterns above are what I scan for in eight seconds. Strip every one of them and you survive the screen. Leave any of them in and you don’t. The 20 minutes of rewriting is the entire job.