ChatGPT Prompts for Recruiter BD: 12 Battle-Tested Prompts Top Agencies Use to Land Clients in 2026

TL;DR

Twelve copy-paste prompts every BD recruiter should have saved in 2026. They cover triage, research, outreach, qualification, and strategy. Each one has structured context, explicit task definitions, and the specific phrases banned from the output - because the difference between AI slop and AI leverage is in the constraints.

The short version

• Structure beats clever phrasing - tagged blocks (CONTEXT, TASK, FORMAT) outperform free-text prompts every time.
• Ban the cliche phrases - explicit "do not use" lists are how you stop output sounding like a recruiter cliche generator.
• Force a specific reference - real LinkedIn post, press line, or hiring signal in the first sentence of any outreach.
• Saved prompts compound - the third time you run a prompt with your ICP filled in, it takes 30 seconds.
• Some prompts deserve to become a tool - hiring signal triage and decision-maker research are now better automated than prompted.

Why Prompts Matter for Recruiter BD

BD is a workflow with thousands of small judgement-and-writing tasks per month. Triage, research, draft, follow up, qualify, refine. AI compresses each of those tasks from 10-30 minutes to 60-90 seconds when the prompt is right. The recruiters winning in 2026 have stopped writing prompts from scratch and started running a library.

What actually changes when you have a prompt library

• Throughput - typical agency BD output rises 1.5-2x within 60 days of consistent prompt use.
• Quality consistency - the worst recruiter on the team writes outreach as good as the best one, because the prompt does the structural lifting.
• Onboarding speed - new recruiters reach productive BD output in 4 weeks, not 4 months.
• Compoundable knowledge - good prompts get refined every quarter and the library appreciates over time.
• Time spent on conversations - the share of recruiter hours on phone calls and meetings rises from ~30% to ~55%.

Anatomy of a High-Output Prompt

Every prompt in this library follows the same structure. It is boring on purpose - the structure is what produces consistency.

1. Role assignment - "You are a senior BD recruiter at a UK staffing agency". One sentence, specific.
2. Tagged context blocks - CONTEXT, ICP, TARGET, COMPANY. Structured, not prose.
3. Task definition - explicit, numbered, with constraints (length, tone, banned phrases).
4. Format specification - markdown table, bullet points, max word count, citation requirements.
5. Negative space - what NOT to do. The "do not use these phrases" line is where most prompts win or lose.

Component	Lazy Version	High-Output Version
Role	"You are a recruiter"	"You are a senior BD recruiter at a UK fintech-focused staffing agency"
Context	Free-text paragraph	Tagged blocks with named fields
Task	"Write a cold email"	"Write a 90-word cold email opening with one specific reference, banning [list]"
Format	No specification	Markdown table, max length, citation rules
Constraints	None	Word count, banned phrases, tone, must-include reference

1. Hiring Signal Triage

When to use it: When you have a list of 30 hiring signals and 20 minutes to decide who to call first.

You are a senior BD recruiter at a UK staffing agency specialising in [VERTICAL]. I have the following hiring signals from the last 7 days:

<SIGNALS>
[PASTE SIGNALS HERE - one per line, format: company | signal type | date | source URL]
</SIGNALS>

<MY_ICP>
- Industry: [INDUSTRY]
- Company size: [HEADCOUNT RANGE]
- Geography: [REGION]
- Role types we place: [ROLE LIST]
- Recent placements at similar companies: [3-5 EXAMPLES]
</MY_ICP>

<TASK>
Triage these signals into three buckets:
1. Call-today: signals where my placement track record + the signal type + the company profile suggest a 30%+ likelihood of a callback this week.
2. Email-this-week: signals worth a single warm email but not a phone call.
3. Skip: signals that look strong on the surface but my ICP says no.

For each signal in buckets 1 and 2, give me a single-sentence reason why I am calling now, and the most likely role to lead with.
</TASK>

<FORMAT>
Markdown table with columns: Company | Bucket | Reason | Lead Role
No more than 1 sentence per cell.
</FORMAT>

Why this works: Cuts triage from 90 minutes to 12. The recruiter still makes the call - the prompt just orders the queue.

2. Decision-Maker Research Brief

When to use it: Before any cold call or LinkedIn DM to a hiring manager you have not spoken to before.

You are a recruitment BD analyst. I need a one-page research brief on a hiring manager I am about to contact.

<TARGET>
Name: [FULL NAME]
Company: [COMPANY]
LinkedIn URL: [URL]
Role title: [TITLE]
</TARGET>

<MY_AGENCY>
We place [ROLE TYPES] for [INDUSTRY] companies in [REGION]. Recent wins include [2 BRIEF CASE STUDIES].
</MY_AGENCY>

<TASK>
Produce a brief covering:
1. Career trajectory in 3 bullets.
2. Likely current priorities based on company stage and recent news.
3. Two or three credible openers grounded in their actual work, not generic flattery.
4. One topic to AVOID (e.g. recent layoff, controversial post, sensitive transition).
5. Most likely tech, sales, or ops role they will hire next based on the team they currently lead.

Cite the source URL for each claim. Anything you cannot source, mark as "inferred" - do not invent.
</TASK>

<FORMAT>
Headed sections with bullet points. No prose paragraphs. Maximum one page.
</FORMAT>

Why this works: Replaces 25 minutes of LinkedIn-stalking with a structured brief. The 'cite sources' line keeps it honest.

3. Cold Email Personalisation

When to use it: When the templated cold email is ready and the personal hook is missing.

You are a senior copywriter who has written cold emails for recruitment agencies for 10 years.

<MY_TEMPLATE>
[PASTE YOUR EXISTING COLD EMAIL TEMPLATE HERE]
</MY_TEMPLATE>

<TARGET>
Name: [NAME]
Company: [COMPANY]
Role: [ROLE]
LinkedIn snippet (most recent activity): [PASTE]
Most recent company news: [PASTE]
</TARGET>

<TASK>
Rewrite the opener (first 2 sentences only) so it:
- References ONE specific, real, citable thing about the target or their company in the first 12 words.
- Does not use any of these phrases: "saw your post", "noticed you recently", "hope this email finds you well", "exciting opportunity".
- Sounds like a peer, not a vendor.
- Pivots naturally into the existing template's value proposition.

Give me 3 alternative openers in different tones: direct, curious, and contrarian.
</TASK>

<FORMAT>
Numbered list. Each opener no more than 35 words. Show the joined sentence with the rest of the template implied.
</FORMAT>

Why this works: The opener is the only part of a cold email that needs personalisation. This prompt isolates it.

4. LinkedIn DM That Does Not Sound Like a Recruiter

When to use it: Inbound or outbound DM where you need to feel like a peer, not a pitch.

You are a recruitment-agency founder who writes LinkedIn DMs that consistently get replies from senior decision-makers.

<TARGET>
Name: [NAME]
Title: [TITLE]
Company: [COMPANY]
Most recent LinkedIn activity (post, comment, or share): [PASTE WITH DATE]
</TARGET>

<MY_ANGLE>
[ONE SENTENCE on what would actually be useful or relevant to this person from your point of view]
</MY_ANGLE>

<TASK>
Write a LinkedIn DM that:
- Opens with a one-line reaction to their actual recent activity, not a flattery line.
- Names the relevance to them in 1 sentence.
- Closes with a low-commitment ask (not "jump on a call", more like "worth me sending a 2-line intro?").
- Avoids the phrases: "I help", "I work with", "I am reaching out", "would love to connect".
- Sounds like 2 humans messaging each other on a Tuesday.

Maximum 70 words total.
</TASK>

Why this works: The 'avoid phrases' list is the trick. AI defaults to recruiter cliche; this filters it.

5. Discovery Call Prep

When to use it: 20 minutes before a first call with a potential client.

You are a senior agency recruiter prepping for a 30-minute discovery call.

<COMPANY>
Name: [NAME]
Recent news: [PASTE]
Public hiring signals: [PASTE]
LinkedIn headcount trend (if known): [PASTE]
</COMPANY>

<HIRING_MANAGER>
[NAME, TITLE, BRIEF BACKGROUND]
</HIRING_MANAGER>

<MY_AGENCY>
We place [ROLE TYPE] for [VERTICAL]. Recent comparable placements: [2 EXAMPLES with outcomes].
</MY_AGENCY>

<TASK>
Produce a discovery-call prep doc covering:
1. Three things to confirm in the first 5 minutes (about the role, the team, the urgency).
2. Five qualifying questions specific to this company - not generic.
3. Two case studies from my own experience that map to their likely pain.
4. One question they will probably ask me, and a 2-sentence answer ready.
5. The "next-step ask" I should aim for if the conversation goes well.
</TASK>

<FORMAT>
Numbered sections, bullet points. Maximum one page. No prose.
</FORMAT>

Why this works: Replaces 30 minutes of fragmented prep with a single page the recruiter can scan in 90 seconds.

6. Funding Round Response

When to use it: Same-day reach-out when a target company announces a funding round.

You are a fast-moving BD recruiter responding to a funding announcement within hours.

<FUNDING_EVENT>
Company: [NAME]
Round: [SERIES + AMOUNT]
Lead investor: [NAME]
Date: [DATE]
Stated use of funds: [PASTE FROM PRESS RELEASE]
</FUNDING_EVENT>

<MY_AGENCY>
We place [ROLE TYPE] for [STAGE/VERTICAL]. Top 3 placements at similar post-funding companies: [BRIEF]
</MY_AGENCY>

<TASK>
Draft three outreach assets:
1. A 60-word email to the CEO referencing the round, the use of funds, and offering one specific way we have helped other companies at the same milestone. No "congrats" opener.
2. A 80-word LinkedIn DM to the most likely hiring manager (CTO if engineering hires, VP Sales if GTM, etc.).
3. A talking script for a 90-second voicemail if they pick up the phone.

Tone: peer, not vendor. Cite the specific use-of-funds line.
</TASK>

Why this works: Funding rounds are time-sensitive. This prompt produces three vehicles in 2 minutes flat.

7. Leadership Change Outreach

When to use it: When a new VP, Director, or C-level lands at a target company.

You are a BD recruiter reaching out to a newly-appointed leader in their first 60 days.

<APPOINTMENT>
New leader: [NAME, TITLE, COMPANY]
Previous role: [WHERE THEY CAME FROM]
Date appointed: [DATE]
Public statement (if any): [PASTE]
</APPOINTMENT>

<MY_AGENCY>
We place [ROLE TYPE] - especially relevant for new leaders building out a team. Recent comparable: [BRIEF].
</MY_AGENCY>

<TASK>
Write a 90-word email that:
- References the appointment without sounding like a press-release scrape.
- Acknowledges the well-documented pattern of new leaders building 3-5 hires in their first 90 days.
- Offers one concrete, agency-specific way to help (e.g. shortlist of 3 pre-vetted candidates, market map, prior team's typical hires).
- Closes with a 2-line ask, not "call me back".

No congrats opener. No "exciting times".
</TASK>

Why this works: New leaders hire fast. This prompt frames the agency as informed and useful, not opportunistic.

8. Lapsed Client Reactivation

When to use it: Reaching out to a client you have not placed with in 6+ months.

You are a relationship-led BD recruiter reactivating a lapsed client without sounding desperate.

<CLIENT_HISTORY>
Company: [NAME]
Last placement: [DATE, ROLE, OUTCOME]
Reason engagement lapsed (if known): [PASTE OR INFER]
Recent public news: [PASTE]
Recent hiring signals: [PASTE]
</CLIENT_HISTORY>

<MY_AGENCY>
[BRIEF - what we have done since, or for them, or for similar companies]
</MY_AGENCY>

<TASK>
Draft a reactivation email that:
- Opens with a specific reference to recent activity at the company - signals, news, hires - not "checking in".
- Acknowledges the gap honestly without apologising.
- Gives one piece of value upfront (a market insight, a candidate name with permission, a benchmark) before any ask.
- Ends with a low-commitment ask: "worth a 15-minute swap on what is changing in [vertical]?"

Maximum 110 words. Tone: warm, professional, slightly informed.
</TASK>

Why this works: Reactivation emails are the highest-ROI outreach an agency does. This prompt prevents the desperate 'just checking in' trap.

9. Cross-Sell to Existing Clients

When to use it: When you currently work one role for a client and want to widen the relationship.

You are an account-led recruiter widening an existing client relationship from one role type to two.

<CURRENT_ENGAGEMENT>
Client: [NAME]
What we currently place for them: [ROLE TYPE]
Performance to date: [BRIEF]
Primary contact: [NAME, TITLE]
</CURRENT_ENGAGEMENT>

<EXPANSION_TARGET>
Role type to introduce: [NEW ROLE TYPE]
Hiring signal that suggests they need this: [PASTE]
Likely decision-maker for the new role: [NAME, TITLE]
</EXPANSION_TARGET>

<TASK>
Draft a 3-step expansion sequence:
1. A 70-word email from me to my current contact asking for a warm intro to the new decision-maker - referencing the signal.
2. A 90-word email I send to the new decision-maker once introduced - leveraging the existing relationship without name-dropping inappropriately.
3. A 1-line follow-up message 5 days later if no reply.

All three messages should be self-contained but connected. Avoid "synergy", "expand", "scale".
</TASK>

Why this works: Existing-client expansion is 3-5x cheaper than new-logo. This sequence runs the play cleanly.

10. Brief Qualification

When to use it: On the call, when you need to decide if this brief is worth working.

You are a brutally honest delivery recruiter qualifying a new brief from a client.

<BRIEF>
Role title: [TITLE]
Company: [NAME]
Salary: [RANGE OR EQUITY DETAIL]
Location: [REMOTE / HYBRID / OFFICE / CITY]
Hiring manager: [NAME, TITLE]
Stated must-haves: [PASTE]
Stated nice-to-haves: [PASTE]
Timeline: [URGENCY]
Comp vs market: [BRIEF VIEW]
Recent failed hires for this role (if any): [PASTE]
</BRIEF>

<MY_DESK>
Average fill rate on similar briefs: [%]
Typical time-to-shortlist: [DAYS]
Capacity this month: [BRIEF]
</MY_DESK>

<TASK>
Score this brief from 1 to 5 across:
1. Fillability (talent supply vs spec)
2. Compensation realism vs market
3. Timeline realism
4. Hiring manager engagement quality
5. Strategic value to my agency

Then give me:
- A go / pause / decline recommendation.
- The two riskiest assumptions in the brief.
- Three pre-emptive questions to ask the hiring manager before committing.
</TASK>

<FORMAT>
Scores in a 5-row table, then prose for the recommendation.
</FORMAT>

Why this works: Most briefs an agency works are unfillable in their stated form. This prompt catches that in 90 seconds.

11. ICP Refinement

When to use it: Quarterly, when you suspect your ICP is too broad or out of date.

You are a BD strategist helping a recruitment agency tighten its ICP based on actual placement data.

<MY_DATA>
Total placements last 12 months: [N]
Top 10 placements by revenue: [LIST WITH COMPANY, ROLE, FEE, TIME-TO-FILL]
Bottom 10 placements by margin or pain: [LIST]
Briefs taken but not filled (last 12 months): [N + REASONS]
Current stated ICP: [PASTE]
</MY_DATA>

<TASK>
Analyse the data and produce:
1. The 3 attributes most predictive of high-revenue, fast-fill placements.
2. The 3 attributes most predictive of failed or low-margin placements.
3. A revised ICP definition (1 paragraph max).
4. 5 firmographic filters I can apply in any BD tool (signals platform, CRM, sales nav).
5. The single biggest blind spot in my current ICP.

Be direct. If my ICP is too broad, say so. If I am chasing the wrong companies, say so.
</TASK>

Why this works: ICP drift is the silent killer of agency BD productivity. Quarterly refinement keeps the desk sharp.

12. Objection Talking-Track

When to use it: Before a renewal or expansion conversation where you know objections are coming.

You are a calm, prepared agency owner walking into a difficult client conversation.

<OBJECTIONS_EXPECTED>
1. [OBJECTION 1 - paste likely client position]
2. [OBJECTION 2]
3. [OBJECTION 3]
</OBJECTIONS_EXPECTED>

<MY_POSITION>
Outcomes delivered to date: [BRIEF]
Comparable market data on fees / outcomes: [BRIEF]
Walk-away point: [WHAT I WILL NOT ACCEPT]
</MY_POSITION>

<TASK>
For each objection, give me:
1. The most likely emotional driver behind it (cost pressure, AI fear, last-bad-experience, etc.).
2. A 2-3 sentence response that acknowledges the objection without conceding the point.
3. The single best question to ask back, to shift the conversation.
4. An honest answer if their objection is actually valid.

End with a one-paragraph "anchoring summary" I can open the meeting with to set the frame.
</TASK>

<FORMAT>
Numbered per objection. Direct, practical, no fluff.
</FORMAT>

Why this works: The 'honest answer if valid' line is what keeps the prompt grounded - AI tends to defend reflexively otherwise.

How to Sound Human, Not AI

The biggest reason AI-assisted outreach fails is that recruiters trust the first draft. The model gets you to 80%; the recruiter takes it the rest of the way. Six rules consistently separate "this sounds like AI" from "this sounds like a thoughtful peer".

• Cut the first sentence - AI almost always opens with throat-clearing. Delete it. Start with the reference.
• Replace adjectives with specifics - "exciting opportunity" becomes "Series A fintech, 3 engineering hires next 30 days".
• Read it aloud - if you would not say it that way on a phone call, do not send it as an email.
• Cut every "I help" or "I work with" - those are the verbal equivalent of a flashing "AI" badge.
• Cut adverbs - "really", "very", "extremely". They are filler that AI loves and humans should not.
• Add one minor imperfection - a parenthetical aside, a contraction, a slightly off-script line. Polish kills outreach.

“Anyone who pastes raw AI output into a cold email is the reason cold emails are dying. The prompt does the structural work. The recruiter does the human work. That has not changed.”

- Felix Hermann, Cofounder @ Boilr

When to Skip the Prompt and Use a Tool Instead

Some of these prompts are genuinely better as a saved-and-run habit. Others are workflows that have outgrown the chat window and should now be a tool. Knowing which is which is what separates recruiters running a real AI stack from recruiters running 47 ChatGPT tabs.

Workflow	Prompt or Tool?	Why
Hiring signal triage	Tool	Runs daily, needs fresh data, scales beyond chat context
Decision-maker research	Tool	Repetitive, benefits from enrichment APIs and integration
Cold email personalisation	Hybrid	Tool generates draft, recruiter polishes
Discovery call prep	Prompt	Per-call, recruiter judgement matters
Brief qualification	Prompt	Recruiter judgement first; tool can score later
ICP refinement	Prompt	Strategic, quarterly, needs recruiter ownership
Objection talking-track	Prompt	Pre-meeting, context-heavy

When Boilr Replaces the Prompt

Boilr automates the prompts that have outgrown the chat window. Hiring signal triage runs 24/7, not when the recruiter remembers to paste signals into ChatGPT. Decision-maker research is enriched against verified data, not inferred from public bios. Outreach personalisation pulls from the actual signal that triggered the engagement, not a generic LinkedIn snippet.

• Always-on signal triage - funding rounds, leadership changes, expansion alerts, hiring velocity, sorted by your ICP automatically.
• Decision-maker identification - the right hiring manager surfaced per signal, with verified contact and trajectory.
• Two-way candidate matching - candidates scored against live demand, not static job descriptions.
• Outreach drafts grounded in signal - the email opens with the actual reason the company is hiring, not a templated hook.
• Free prompt library - the prompts in this article live at boilr.ai/prompts-for-recruiters, regularly refreshed.

Manual prompts vs Boilr automation - honest comparison

When prompts win

• Strategic, low-volume tasks - ICP refinement, objection prep, brief qualification
• Recruiter judgement matters most - the prompt amplifies, does not replace
• Per-call prep - context shifts every time
• Free, no setup - any recruiter, any agency, any tier

When automation wins

• High-volume, repetitive workflows - signal triage, research, decision-maker ID
• Always-on coverage - signals do not wait for the recruiter to open ChatGPT
• Verified data sources - enrichment beats inference
• Team consistency - the whole desk runs the same playbook

Prompts vs Automated Tools - Decision Framework

A simple framework for deciding which workflows belong in a prompt library and which belong in a tool. The shorthand is volume and repeatability - high volume + high repeatability = tool, low volume + high context = prompt.

1. Run the workflow weekly? If yes, lean towards a tool. If no, prompt is fine.
2. Does it need fresh external data? If yes (signals, news, hiring data), tool. If no, prompt.
3. Does the team need to run it consistently? Tool. Prompts drift across team members.
4. Is recruiter judgement the main input? Prompt. Tools cannot replace context.
5. Is the output time-sensitive? Tool - automation runs 24/7.

Frequently Asked Questions

What makes a ChatGPT prompt good for recruiter BD?

Three things separate prompts that work from prompts that produce slop. First, role assignment - tell the model exactly who to be (a senior BD recruiter at a UK staffing agency, not just 'an assistant'). Second, structured context using tagged blocks (CONTEXT, TASK, FORMAT) so the model knows what is information and what is instruction. Third, explicit constraints on length, tone, and forbidden phrases - this is where the difference between recruiter-cliche output and genuinely usable copy lives.

Should I use ChatGPT, Claude, or Gemini for these prompts?

All three handle these prompts well in 2026. Claude tends to produce slightly more grounded copy with better instruction-following on long prompts. ChatGPT (GPT-5) is fast and strong on the structured-research prompts. Gemini is fine for short-form outreach. The bigger variable is whether you give the model good context, not which model you choose.

How long do these prompts take to actually save time?

After three or four runs you will have the placeholders filled in for your typical role types and ICPs, and the prompts run in under a minute. The first time you use one, expect 5-10 minutes of context entry. After that, the prep work compounds - your decision-maker brief prompt becomes a 30-second task per target.

Do these prompts work without ChatGPT Plus or Claude Pro?

Yes. Free tiers handle these prompts. The paid tiers help with longer context windows, faster responses, and better handling of multi-step prompts. If your agency runs more than 50 BD touches a week, paid tiers pay for themselves quickly through speed alone.

How do I stop AI-written outreach sounding like AI?

Three rules. First, ban the cliche phrases explicitly in the prompt - 'I hope this email finds you well', 'I work with', 'exciting opportunity'. Second, force a specific reference in the first sentence (a real LinkedIn post, a press line, a hiring signal). Third, edit the output - the model gets you to 80%, the recruiter takes it home. Anyone who pastes raw AI output into a cold email is the reason cold emails are dying.

Are there compliance issues with using ChatGPT for recruiter BD?

Two areas to think about. GDPR - if you paste candidate or client personal data into a public model, you are processing it. For most agencies, this means avoiding sensitive personal data in free-tier ChatGPT and using the enterprise or workspace tiers that do not train on inputs. The EU AI Act adds obligations once AI is making or substantially supporting hiring decisions - drafting outreach is generally fine, automated screening of candidates is not. Check your DPIA before scaling.

Should I save my prompts in ChatGPT or in a separate tool?

Save them somewhere outside the chat tool. ChatGPT custom instructions are useful for recurring context, but a dedicated prompt library (Notion, a shared doc, or a tool like PromptHub) gives you version control, sharing across the team, and the ability to refine prompts over time. The agencies winning with prompts treat them like product - they iterate.

What is the most common prompt mistake recruiters make?

Vague task definitions. 'Write me a cold email to this person' produces generic output. 'Write a 90-word cold email to this person, opening with a one-sentence reference to their post about X, avoiding the phrases [list], and closing with a 2-line ask not a meeting request' produces something usable. The more specific the task, the better the output.

Can I just use a tool like Boilr instead of running these prompts manually?

For some of these workflows, yes. Boilr automates the hiring signal triage, decision-maker research, and outreach personalisation steps so the prompts run in the background while you focus on conversations. For the more strategic prompts - ICP refinement, objection handling, brief qualification - manual prompting is still the better tool. The right answer is usually 'both': the tool runs the high-volume prompts, the recruiter runs the high-context ones.

Do these prompts work for in-house TA, not just agency recruiters?

Most do, with minor tweaks. Replace 'agency' with 'TA team' and 'client' with 'hiring manager' in the role assignment. The brief qualification, ICP refinement, and objection handling prompts are particularly useful for in-house teams pushing back on bad briefs internally.

How often should I refresh my prompts?

Every 8-12 weeks for the outreach prompts (because cliche phrases drift and what 'sounds AI' changes), and quarterly for the strategy prompts. Anything you have not touched in 6 months is probably underperforming. Keep a 'last edited' date on every prompt in your library.

What about voice and SMS - do these prompts work there?

The discovery prep, signal triage, and decision-maker research prompts translate directly. The outreach prompts (email, LinkedIn DM) need shortening - voicemail and SMS need 30 words, not 90. Add a length constraint and force-cut adjectives. Voice in particular benefits from a 'say this aloud' filter on the prompt output.

Should I share my best prompts with the rest of the team?

Yes - and most agencies that have done this saw a step-change in BD output. Treat prompts like a sales playbook. The recruiter who refines a great cold-email prompt should publish it to the team library, with an example output. The best agencies in 2026 run weekly 30-minute prompt reviews.

Can I see Boilr's prompt library?

Yes - boilr.ai/prompts-for-recruiters has the live, regularly-updated library, and the in-product prompt builder lets you assemble prompts step-by-step against your own ICP. It is free to use without a Boilr account.

Sources

1. OpenAI - Prompt engineering guide
2. Anthropic - Claude prompt engineering documentation
3. DemandSage - AI Recruitment Statistics 2026
4. Pin.com - The Complete 2026 AI Recruiting Guide
5. Cadient - AI-Driven Recruitment in 2026
6. boilr.ai - Live prompt library and prompt builder