Asset 20 8 2
Does AI recommend your business? Run the free check →

Join 15,000 business owners, marketers and entrepreneurs. The Sunday newsletter you'll be annoyed only arrives once a week.

Article

Claude Prompts for Recruiters That Save Hours

If you are skim reading
The short version: Claude is better than ChatGPT at long, messy CV screening because it handles bigger documents and sticks to instructions without drifting, but it still can't judge whether someone's good at the job.

The short version: Claude is better than ChatGPT at long, messy CV screening because it handles bigger documents and sticks to instructions without drifting, but it still can't judge whether someone's good at the job. Use these prompts to cut admin time on job ads, screening, and candidate outreach, then spend the hours you've saved on the human conversations that decide who you hire.

Why recruiters are moving some of their workflow to Claude

I use both Claude and ChatGPT for different jobs, and for recruitment admin specifically, Claude tends to win on one thing: it holds more context. If you're pasting in 15 CVs plus a job spec plus your own notes on a hiring manager's preferences, Claude handles that volume without losing the thread halfway through. ChatGPT can do it too, but I find it starts summarising rather than following instructions once the input gets long.

That matters for recruiters because the actual bottleneck isn't writing the job ad. It's the hours spent reading CVs that were never going to get an interview. A recruiter handling 40 applications for a mid-level marketing role might spend six or seven hours just on first-pass screening. Claude won't do your judgement for you, but it will do the sorting in minutes rather than hours.

The uncomfortable bit nobody likes saying out loud

Here's what most people writing about AI prompts for recruiters won't tell you plainly: if you feed Claude a CV and a job description and ask it to score the candidate, it will produce a confident-sounding number, and that number is not a hiring decision. It's a pattern match against keywords and structure. A brilliant candidate who wrote a plain, undecorated CV will score lower than a mediocre one who knows how to mirror the job spec's language back at it, because that's literally what both models are trained to notice. If you let the AI score replace your read of the person, you'll filter out exactly the kind of candidate who doesn't play the keyword game but would be excellent at the actual job. Use it to shortlist faster, never to reject automatically.

Prompt 1: Turning a messy job spec into a real job ad

Hiring managers rarely hand you a clean brief. You usually get a bullet list, a salary range, and a half-formed idea of "someone who gets stuff done." Here's a prompt that works:

  • "Here is a rough internal brief for a role: [paste notes]. Write a job ad of around 350 words, in a tone that's direct and specific rather than corporate. Include a clear list of must-have requirements separate from nice-to-haves, a realistic day-one task, and the salary range stated as [insert range]. Avoid vague phrases like 'fast-paced environment' and 'wear many hats.'"

Say you run a 12-person digital agency and need to hire a part-time paid media buyer. You'd paste in three bullet points from the founder, the budget they manage (say £40,000 a month across clients), and the tools they'll use (Meta Ads Manager, Google Ads). Claude will draft an ad that names the tools and the budget instead of saying "experience with digital advertising platforms," which is the kind of vague line that gets you 200 unqualified applicants instead of 20 good ones.

Prompt 2: First-pass CV screening with a strict scoring rubric

This is the one that saves the most time, and it only works if you give Claude a rubric rather than asking it to "review this CV."

  • "Here is the job description: [paste]. Here is a candidate CV: [paste]. Score this candidate out of 10 on three separate criteria: relevant experience, evidence of the specific skills listed as must-haves, and career progression. Give a one-line reason for each score. Do not give an overall recommendation, just the three scores and reasons."

Notice it deliberately doesn't ask for a yes/no recommendation. That's the step you keep for yourself. Run this across 30 CVs and you can build a simple table in about 20 minutes that would normally take most of an afternoon, then you read the top 10 yourself.

If you want a template built specifically for the HR side of this rather than agency recruiting, the ChatGPT prompts for HR post has copy-and-paste versions for small teams doing this without a dedicated recruiter.

Prompt 3: Writing outreach messages that don't read like templates

Recruiters send a lot of cold LinkedIn messages, and most of them are obviously generated or obviously copy-pasted, which candidates can spot in about two seconds. Try this instead:

  • "Write a LinkedIn outreach message to a passive candidate. Their current role is [title/company], and the open role is [paste summary]. Mention one specific, plausible reason this move would be a step up, not a vague compliment. Keep it under 80 words and don't use the word 'exciting.'"

The "don't use the word exciting" instruction isn't a joke. Every AI-generated recruiter message defaults to "exciting opportunity" within the first line, and candidates have started treating that phrase as a signal to ignore the message. Banning specific words in your prompt is a useful technique that most guides skip entirely.

Prompt 4: Interview question sets tied to the actual job spec

Generic interview questions get generic answers. Ask Claude to tie questions directly to the must-haves in your job spec:

  • "Based on this job description [paste], write 8 interview questions. For each must-have requirement, write one behavioural question that would reveal whether the candidate has real experience versus theoretical knowledge. Add a one-line note on what a weak answer sounds like versus a strong one."

That last instruction is the part people miss. Without it, you get questions but no sense of what to listen for, and new or junior interviewers end up scoring confident answers higher than honest ones.

Work with me

Want AI doing the heavy lifting in your marketing?

I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.

Prompt 5: Rejection emails that don't sound robotic or legally risky

Most rejection templates are either so warm they're meaningless or so stiff they sound automated. A workable middle ground:

  • "Write a rejection email for a candidate who reached final interview stage but wasn't selected. Keep it to 100 words. Don't give specific reasons for the decision, avoid saying 'we encourage you to apply again' unless that's likely, and end with a clear, final tone rather than an open-ended one."

Candidates who've reached final stage remember rejections. A vague, falsely encouraging email damages your employer brand more than a short, honest one does.

A step-by-step way to build your own prompt library

  1. Pick the five recruitment tasks that eat the most of your week. For most small agencies that's screening, outreach, job ads, interview prep, and rejections.
  2. Write one prompt per task and save it in a plain document, not scattered across chat history you'll never find again.
  3. Test each prompt on a real but anonymised example and tweak the wording until the output needs minimal editing.
  4. Add constraints, not just instructions: word limits, banned phrases, required structure. Constraints are what separate a usable prompt from a vague one.
  5. Review the prompts every quarter, because job market language and your own hiring needs shift faster than people expect.

If you're the candidate side of this equation rather than the recruiter, the same logic applies in reverse, and there's a detailed breakdown of Claude prompts for job applications and cover letters worth reading so you understand exactly what the AI on the other side of the table is likely doing with a CV.

Where this fits with the rest of your hiring process

None of this replaces real screening tools or a trained eye for culture fit. Think of Claude as the first filter, not the final one. If you're screening at volume, it's worth pairing these prompts with dedicated software rather than doing everything inside a chat window, and the piece on AI tools for recruiters to screen candidates faster covers where that line sits.

And if you manage hiring for a small business without an internal recruiter, it's worth reading how the same prompts apply to resumes from the other direction, covered in how to write better resumes using AI prompts, so you can spot when a CV has clearly been run through the same kind of tool you're using to screen it.

Frequently asked questions

Is Claude better than ChatGPT for recruiters?

For long documents and strict instruction-following, such as screening multiple CVs against a rubric, Claude tends to perform more consistently. For quick outreach drafts or brainstorming, either works fine, so it comes down to the task rather than one being universally better.

Can Claude replace a human recruiter?

No. It can speed up screening, drafting, and admin, but it can't judge cultural fit, read between the lines in an interview, or take responsibility for a hiring decision. Use it for the first pass, not the final call.

Is it risky to use AI scoring to reject candidates automatically?

Yes, and this is worth taking seriously. AI scoring tends to reward candidates who mirror the job spec's exact wording, which can unfairly filter out strong candidates who write plainly. Use AI scores to shortlist faster, never as an automatic rejection tool.

What's the single most useful Claude prompt for recruiters?

A CV screening prompt that asks for separate scores on experience, must-have skills, and career progression, with one-line reasons for each, rather than a single overall recommendation. It keeps the final judgement with you, where it belongs.

Run a business, a brand or a blog? I feature tools, products and expert guest posts for an audience of 15,000. See how it works

Published and maintained by the Lilach Bullock team, covering marketing, AI and business growth.
Your buyers are asking AI who to use. Does it say you?

See for free whether ChatGPT, Claude, Perplexity, Gemini and Google name you, and get the plan to become the answer.

Check my AI visibility →
Sundays only

Get the Sunday newsletter.

One email a week. AI experiments, marketing tactics, and the workflows Lilach is building right now in her own business.

Subscribe free

Let’s get your marketing running on AI.

Book a free 30-minute call

We figure out what you need, where AI fits in, and what working together would look like.

Book the call →

Or take the 30-second calculator

You’ll see the hours and the money quietly leaking out of your week, and the three workflows worth building first.

Take the calculator →

Or grab the free AI resource library

Prompt packs, templates, checklists, and swipe files. The exact tools I build for paying clients. Yours, free.

Get the library →
Keep reading

More from the blog.