50 Claude Prompts That Run Your Recruiting Process

You are a hiring process designer who builds resume rubrics that survive an audit. <context> Three of us screen resumes for the same role and we disagree constantly. I want one rubric so our decisions match. </context> <inputs> - Role, level, and team: [DETAILS] - The three outcomes this hire owns: [OUTCOMES] - Must-have qualifications: [LIST] - Signals we value that are not on the must-have list: [LIST] - Volume of applications per week: [NUMBER] </inputs> <task> Build a screening rubric with four to six criteria, each scored 1 to 4 with a written anchor describing what a 1, 2, 3, and 4 look like on a resume. Add a pass threshold, a rule for handling non-traditional backgrounds, a list of things reviewers must ignore, and a 90-second review order so screening stays fast. </task> <constraints> - Anchors must reference evidence visible on a resume, not inferred personality. - Explicitly exclude school prestige, employment gaps, photos, names, and graduation years from scoring. - The whole rubric must fit on one page. </constraints> <format> A scoring table with anchors, then the threshold rule, then the ignore list, then the 90-second review order. </format>

You are an interviewer who can qualify a candidate properly in 25 minutes. <context> My phone screens ramble and I still end up unsure. I need a tight question set that produces a clear advance or reject. </context> <inputs> - Role and level: [DETAILS] - The two riskiest assumptions about candidates for this role: [RISKS] - Must-confirm logistics: [COMP, LOCATION, NOTICE, VISA, HOURS] - What the next interview stage tests: [NEXT STAGE] - Deal-breakers: [LIST] </inputs> <task> Write a 25-minute phone screen: a two-minute framing script, six core questions that test my riskiest assumptions with one follow-up probe each, the logistics block, a candidate question window, and a decision rule that tells me what an advance actually requires. Include what a weak, adequate, and strong answer sounds like for each core question. </task> <constraints> - Do not ask questions the resume already answers. - No brainteasers and no hypotheticals with no right answer. - Never ask current or past salary; ask about expectations against our band. </constraints> <format> A timed guide with questions, probes, and answer benchmarks, then a five-line scorecard I fill in before leaving the call. </format>

You are an industrial-organisational psychologist who designs structured interviews. <context> Our interviews are unstructured conversations, so every interviewer forms an opinion from different evidence. I need one guide per competency that anyone on the panel can run. </context> <inputs> - Role, level, and team: [DETAILS] - Competency this interview covers: [COMPETENCY] - Interview length: [MINUTES] - Interviewer experience level: [FIRST TIME OR TRAINED] - Real situations from the job this competency shows up in: [EXAMPLES] </inputs> <task> Build a structured interview guide: a scripted opening, three main questions grounded in real job situations, two probes per question that push for specifics, the evidence each question is meant to collect, a 1 to 4 rating scale with behavioural anchors, and a note-taking template that separates what the candidate said from what the interviewer concluded. </task> <constraints> - Same questions in the same order for every candidate. - Probes must ask what the candidate personally did, not what the team did. - No questions touching protected characteristics or personal life. </constraints> <format> A ready-to-run guide with timings, question blocks, anchors, and the split note-taking template. End with three answers that sound impressive but are actually weak, and what to probe next. </format>

You are an interview content designer building a shared question library. <context> Our interviewers reuse the same three questions and candidates have memorised the answers. I want a bank of behavioural questions per competency with real probes. </context> <inputs> - Competencies we hire against: [LIST OF 4 TO 6] - Role family and level: [DETAILS] - Our culture in three honest words: [WORDS] - Questions we already overuse: [LIST] - Situations candidates will genuinely face here: [EXAMPLES] </inputs> <task> For each competency, write four behavioural questions, each with two probes, the specific evidence to listen for, and one common rehearsed answer plus how to get past it. Then flag which questions are safe for a first-round interviewer and which need a senior interviewer to score fairly. </task> <constraints> - Questions must ask about something that already happened, not a hypothetical. - Avoid questions where the "right" answer is obvious. - Do not reuse any question from my overused list. </constraints> <format> Group by competency. For each question show Question, Probes, Listen for, Rehearsed answer trap, Interviewer level. End with a rotation plan so questions do not leak. </format>

You are an assessment designer who builds work samples candidates do not resent. <context> I want to test real ability without asking for free labour or eight unpaid hours. I need a work sample brief plus the rubric to grade it. </context> <inputs> - Role and level: [DETAILS] - The core skill I need proof of: [SKILL] - Maximum time the exercise should take: [HOURS] - Tools the candidate may use: [LIST, INCLUDING AI POLICY] - A realistic but non-confidential scenario: [SCENARIO] - Who will grade it: [GRADERS] </inputs> <task> Write the candidate-facing brief with scenario, deliverable, explicit time cap, what we will and will not evaluate, and our policy on AI assistance. Then write the grader-facing rubric with three to five criteria, 1 to 4 anchors, and a calibration example of a passing and a failing submission. Add a 15-minute follow-up discussion guide about their submission. </task> <constraints> - The exercise must not be work we would otherwise pay for. - Stay inside the time cap; cut scope until it fits. - Grading must be possible in under 20 minutes per submission. </constraints> <format> Two documents: Candidate Brief, then Grader Rubric with calibration examples, then the follow-up discussion questions. </format>

You are a technical recruiter who can hold a credible conversation about a stack you do not code in. <context> I recruit for a technical role but I am not an engineer. I need a cheat sheet so I can screen properly and not waste the hiring manager time. </context> <inputs> - Role and stack: [ROLE PLUS TECHNOLOGIES] - Level we are hiring: [JUNIOR, MID, SENIOR, STAFF] - What this person will build: [PROJECT CONTEXT] - Adjacent technologies that count as transferable: [LIST] - Hiring manager must-haves: [LIST] </inputs> <task> Write a recruiter cheat sheet: what each technology in the stack actually does in plain language, the difference between someone who used it and someone who owns it, five screening questions I can ask and grade without being an engineer, the answers that signal depth, the answers that signal surface-level exposure, and the resume phrases that overstate experience. </task> <constraints> - No jargon unless you define it in the same sentence. - Questions must be gradeable by a non-engineer using your answer keys. - Include how to tell a senior from a mid-level answer. </constraints> <format> Sections: Stack in plain English, Depth versus exposure, Five questions with answer keys, Resume phrases to probe. Keep it to two pages. </format>

You are an interview designer who builds live case exercises for commercial and strategic roles. <context> I need a live case exercise that shows how a candidate thinks under mild pressure, run inside a normal interview slot. </context> <inputs> - Role and level: [DETAILS] - The judgement I need to observe: [PRIORITISATION, ANALYSIS, TRADE-OFFS, STAKEHOLDERS] - A real business situation, anonymised: [SITUATION] - Data I can hand them: [WHAT THEY GET] - Slot length: [MINUTES] - Interviewers in the room: [WHO] </inputs> <task> Design the exercise: candidate-facing prompt, the data pack summary, a minute-by-minute run of show, three curveballs the interviewer introduces at set moments, and a rubric scoring structure of thinking, use of data, handling of ambiguity, and communication with 1 to 4 anchors. Add the two failure modes interviewers cause themselves and how to avoid them. </task> <constraints> - The case must have no single correct answer; score reasoning, not the conclusion. - Every candidate gets the same curveballs at the same time. - Keep the data pack small enough to absorb in five minutes. </constraints> <format> Candidate prompt, run of show table, curveball scripts, rubric with anchors, interviewer pitfalls. </format>

You are a hiring process architect who removes redundant interview stages. <context> Our loop has too many stages, they overlap, and good candidates drop out. I need a redesigned loop that tests each thing exactly once. </context> <inputs> - Current stages and who runs them: [LIST] - Total elapsed time from application to offer today: [DAYS] - Competencies we must assess: [LIST] - Where candidates drop out: [STAGE] - Constraints: [PANEL AVAILABILITY, COMPLIANCE, REMOTE] </inputs> <task> Redesign the loop. Map each competency to exactly one stage, show which current stages are redundant and should be merged or cut, assign an owner and a time box to each stage, define the advance criteria between stages, and give a target elapsed time. Then show the same loop compressed into a single onsite day for candidates with competing offers. </task> <constraints> - No competency assessed twice unless you justify the duplication. - Total candidate time investment must go down, not up. - Every stage needs a named decision it produces. </constraints> <format> A coverage matrix of competencies against stages, then the stage-by-stage plan with owners and time boxes, then the compressed version, then what we lose by compressing. </format>

You are a seasoned interviewer who knows the difference between a red flag and a bias. <context> Something in a candidate profile or interview is nagging at me. I want to probe it properly instead of quietly downgrading them. </context> <inputs> - Role and level: [DETAILS] - The concern, stated plainly: [WHAT IS BOTHERING ME] - Evidence I actually have: [FACTS ONLY] - What I am inferring from that evidence: [MY ASSUMPTION] - Stage we are at: [SCREEN, ONSITE, REFERENCES] </inputs> <task> First, classify my concern as a legitimate job-related risk, a bias dressed as a concern, or insufficient evidence, and explain which. If it is legitimate, write three probing questions that test the risk directly plus what a reassuring and a confirming answer sound like. If it is bias, name the bias and give me the reframe. Then list the concerns I am legally and ethically required to drop. </task> <constraints> - Job gaps, career changes, non-linear paths, and accents are not red flags by default. - Every probe must connect to a job requirement I can point to. - Say clearly when the honest answer is "you do not have enough evidence yet". </constraints> <format> Classification with reasoning, then the probe set with answer interpretations, then the drop list. </format>

You are a reference checker who gets useful signal instead of polite nothing. <context> Reference calls always turn into "they were great to work with". I need a guide that surfaces real information from a 20-minute call. </context> <inputs> - Candidate name and role we are hiring for: [DETAILS] - Referee relationship: [MANAGER, PEER, DIRECT REPORT, CLIENT] - The two open questions from our interviews: [WHAT WE ARE UNSURE ABOUT] - Strengths the candidate claimed: [CLAIMS TO VERIFY] - Jurisdiction rules I must respect: [REGION] </inputs> <task> Write a 20-minute reference guide: an opener that sets expectations and gets the referee talking, verification questions for the candidate claims, two questions targeting our open concerns, a calibration question that forces a comparison against other people at that level, a question about how to manage this person well, and a closing question that invites anything unsaid. Add how to read hesitation, over-praise, and deflection. </task> <constraints> - No questions about health, family, age, or anything outside job performance. - Ask for specific examples, not ratings. - Note where a reference is legally allowed to confirm only dates and title. </constraints> <format> The call guide with questions and follow-ups, then a signal interpretation section, then a short written summary template for the hiring file. </format>

You are a hiring strategist who writes scorecards that define success before anyone interviews. <context> We are opening a role and nobody has agreed what good looks like. I need a scorecard the hiring manager and panel sign off on before sourcing starts. </context> <inputs> - Role, level, team, and manager: [DETAILS] - Why this role exists now: [BUSINESS REASON] - What must be true 12 months in for this hire to be a win: [OUTCOMES] - Competencies the work demands: [LIST] - Constraints: [BUDGET, LOCATION, TEAM MATURITY] </inputs> <task> Write a role scorecard with: the mission in one sentence, three to five measurable outcomes with target and timeframe, the competencies required to hit each outcome, the behaviours that would sink this role on this team, and the panel assignment showing which interviewer assesses which competency. End with the three questions the hiring manager must answer before this is approved. </task> <constraints> - Outcomes must be measurable and time-bound; no "improve collaboration". - Every competency must trace back to at least one outcome. - Maximum one page. </constraints> <format> Mission, Outcomes table with metric and timeframe, Competency map, Anti-patterns, Panel assignment, Open questions for sign-off. </format>

You are an assessment specialist who converts opinions into comparable scores. <context> My panel gives ratings that mean different things to each of them. I need anchored scales so a 3 means the same thing to everyone. </context> <inputs> - Role and level: [DETAILS] - Competencies being scored: [LIST] - Scale we use: [1 TO 4 OR 1 TO 5] - Level expectations at our company: [WHAT SENIOR MEANS HERE] - Examples of past strong and weak hires: [SHORT DESCRIPTIONS] </inputs> <task> For each competency, write behavioural anchors for every point on the scale, describing observable evidence rather than impressions. Add the one thing interviewers most often misjudge in that competency, a rule for scoring "no evidence collected" separately from "poor performance", and a worked example showing an answer and the score it earns with reasoning. </task> <constraints> - Anchors must be about what the candidate did or said, never about likeability or polish. - Include an explicit "insufficient evidence" option. - No overlapping anchor language between adjacent scores. </constraints> <format> One table per competency with Score, Anchor, Typical evidence. Then the misjudgement notes and two worked examples. </format>

You are a facilitator who runs hiring debriefs where the loudest voice does not win. <context> Our debriefs turn into the most senior person announcing a decision. I need a script that gets independent evidence on the table first. </context> <inputs> - Role and candidate: [DETAILS] - Panel members and what each assessed: [LIST] - Meeting length: [MINUTES] - Decision rule we use: [CONSENSUS, HIRING MANAGER DECIDES, MAJORITY] - Known dynamics: [WHO DOMINATES, WHO STAYS QUIET] </inputs> <task> Write a facilitation script: a pre-meeting instruction that scores must be submitted before anyone talks, an opening that states the scorecard and the decision rule, a round-robin order that starts with the most junior interviewer, the exact prompts to pull evidence rather than opinion, how to handle a split panel, how to handle "I just did not feel it", and a close that states the decision, the reasoning, and the owner of the next step. </task> <constraints> - No scores may change based on hearing someone else score first. - Every claim must be answered with "what did they say or do that shows that". - Timebox each speaker and hold it. </constraints> <format> Pre-work instruction, timed agenda with facilitator lines in quotes, split-panel protocol, closing script, and a decision record template. </format>

You are a decision analyst helping a panel choose between finalists. <context> I have several finalists and the panel keeps comparing them on different dimensions. I need one matrix that makes the trade-offs explicit. </context> <inputs> - Role and scorecard outcomes: [OUTCOMES] - Finalists with a short summary each: [CANDIDATE NOTES] - Panel scores per competency: [SCORES OR NOTES] - Constraints: [COMP EXPECTATIONS, START DATE, VISA, LOCATION] - What the team most needs right now: [PRIORITY] </inputs> <task> Build a comparison matrix scoring each finalist against the scorecard competencies plus the practical constraints. Weight the competencies against what the team needs now and show the weighting explicitly. Then write the honest trade-off for each candidate in two lines, name the risk we accept by choosing each one, and give a recommendation with the conditions that would change it. </task> <constraints> - Do not average away a critical gap; flag any below-threshold score. - Separate "would be great in 18 months" from "can do the job now". - Say clearly where the evidence is too thin to compare. </constraints> <format> Weighted matrix table, per-candidate trade-off summary, risk accepted per choice, recommendation with change conditions. </format>

You are a hiring quality reviewer who turns vague interview notes into usable evidence. <context> An interviewer submitted feedback that is all impression and no evidence. I need to rewrite it into something defensible before the debrief. </context> <inputs> - The feedback as submitted (paste it): [FEEDBACK] - Competency this interview was supposed to assess: [COMPETENCY] - Questions the interviewer asked: [QUESTIONS] - Score given: [SCORE] - Role and level: [DETAILS] </inputs> <task> Separate the feedback into observed evidence, interpretation, and unsupported opinion, quoting the original. Rewrite it as structured feedback that states what the candidate said or did, what it demonstrates against the competency, and what remains unknown. Flag any language that signals bias or references a protected characteristic. Then list the follow-up questions needed to fill the evidence gaps. </task> <constraints> - Do not invent evidence that is not in the original notes. - If a score is unsupported by evidence, say so plainly. - Keep the rewritten feedback under 200 words. </constraints> <format> Three-column split of Evidence, Interpretation, Opinion. Then the rewritten feedback. Then bias flags and the evidence gaps. </format>

You are a hiring manager who writes decision memos that hold up six months later. <context> We have picked a candidate and I need to document the decision for the approver, the panel, and my future self. </context> <inputs> - Role, level, and team: [DETAILS] - Chosen candidate and their score profile: [SUMMARY] - Runner-up and why not them: [SUMMARY] - Known gaps in the chosen candidate: [GAPS] - Comp and start date proposed: [DETAILS] - Approver and what they care about: [WHO] </inputs> <task> Write a one-page decision memo: the recommendation up front, the evidence from the loop that supports it mapped to the scorecard, the two strongest arguments against the decision and why we are proceeding anyway, the onboarding actions that de-risk the known gaps, the comp proposal with justification against band and internal equity, and what we will check at 30, 60, and 90 days to know if we were right. </task> <constraints> - Lead with the decision, not with the process narrative. - Include the counter-argument honestly; a memo with no risks is not credible. - No candidate detail unrelated to the job. </constraints> <format> Recommendation, Evidence against scorecard, Case against, Gap mitigation, Comp rationale, 30/60/90 checkpoints. </format>

You are a fairness reviewer auditing a hiring file for bias before a decision is made. <context> Before we finalise this hire I want a check on whether our written record is about the job or about how much we liked people. </context> <inputs> - Role and scorecard: [DETAILS] - All interview notes and scores, pasted: [NOTES] - Candidate pool composition at each stage if known: [NUMBERS] - Who interviewed whom: [PANEL MAP] - Our decision rule: [RULE] </inputs> <task> Audit the file. Identify language patterns that indicate affinity bias, similarity bias, halo or horn effects, or shifting standards between candidates. Show where the same behaviour was described positively for one candidate and negatively for another. Flag any note referencing appearance, accent, school prestige, family, age, or "culture fit" without a defined behaviour. Then recommend the specific remediation: re-score, collect more evidence, or proceed. </task> <constraints> - Quote the exact wording you are flagging. - Distinguish a genuine job-related concern from a biased framing of the same concern. - Do not conclude bias from a single word without context. </constraints> <format> Findings table with Quote, Pattern, Severity, Recommended action. Then an overall verdict and the three process fixes that would prevent a repeat. </format>

You are a recruiting partner running a proper intake meeting instead of taking an order. <context> A hiring manager wants to open a role. I need an intake structure that gets me everything before I start sourcing, in 45 minutes. </context> <inputs> - Role and team: [DETAILS] - Manager name and how they hire: [FAST, PICKY, ABSENT] - Approved budget and headcount status: [DETAILS] - Deadline or business driver: [WHEN AND WHY] - What I already know: [BACKGROUND] </inputs> <task> Write a 45-minute intake agenda with the exact questions to ask about outcomes, must-haves versus nice-to-haves, the profile of people already succeeding on the team, comp band and flexibility, panel availability, and decision speed. Include the three questions managers resist and how to ask them, the commitments I need in writing before sourcing, and a one-page intake summary I send back for confirmation. </task> <constraints> - Push every wish-list item into must-have or nice-to-have; nothing stays in the middle. - Get panel availability and feedback turnaround committed in the meeting. - Do not leave without a comp band, even a rough one. </constraints> <format> Timed agenda with questions, resistance handling scripts, commitments checklist, then the intake summary template. </format>

You are a levelling architect who defines what each seniority level actually means. <context> We argue about whether a candidate is mid-level or senior because we have never written the levels down. I need definitions our panel can score against. </context> <inputs> - Role family: [FAMILY] - Levels we use: [LIST OF LEVELS] - Competencies that matter: [LIST] - Two people on the team at each level, described: [EXAMPLES] - Company size and stage: [CONTEXT] </inputs> <task> Write level definitions for each competency across each level, describing scope of ownership, degree of ambiguity handled, and impact radius. Then produce the interview signals that distinguish each level from the one below it, the most common levelling mistake for this role family, and a short set of questions that reliably separate adjacent levels. </task> <constraints> - Define levels by scope and ambiguity, not by years of experience. - Each level must be distinguishable in a 45-minute interview. - Keep the language specific to our stage, not generic big-company ladders. </constraints> <format> A grid of competencies against levels, then per-boundary distinguishing signals, then the levelling questions and the common mistake. </format>

You are a facilitator running a calibration session so a panel scores the same answer the same way. <context> My interviewers score inconsistently. I want to run a 60-minute calibration workshop with real material, not a slide deck about bias. </context> <inputs> - Role family and competencies: [DETAILS] - Number of interviewers and their experience: [DETAILS] - Our rating scale and anchors: [SCALE] - Three real anonymised candidate answers, one weak, one borderline, one strong: [ANSWERS] - Time available: [MINUTES] </inputs> <task> Build the workshop: a run of show with timings, the independent scoring exercise using the three sample answers, the reveal and discussion prompts that surface why scores diverged, the resolution rules for disagreement, a short practice round of writing evidence-based feedback, and a set of takeaway commitments each interviewer signs. Include the facilitator answer key with the defensible score for each sample. </task> <constraints> - Interviewers score silently before any discussion. - Spend most of the time on the borderline answer; that is where drift lives. - End with behaviour commitments, not general awareness. </constraints> <format> Run of show table, exercise instructions, facilitator answer key with reasoning, discussion prompts, commitment sheet. </format>

You are a compensation partner who explains a package so the candidate understands its real value. <context> Our offer is competitive but complicated, and candidates only hear the base salary. I need a framing document before the offer call. </context> <inputs> - Role, level, and band: [DETAILS] - Base, bonus, equity, and vesting: [NUMBERS] - Benefits with actual cash value: [LIST] - Candidate current package as they described it: [DETAILS] - Competing offer situation: [KNOWN OR SUSPECTED] - Internal equity constraints: [WHAT I CANNOT BREAK] </inputs> <task> Produce a framing doc: total package broken down by component with a plain-language explanation of each, a year-one and year-four view, the honest comparison against their current package including what they give up, the three components candidates undervalue and how to explain each in one sentence, and the two questions this candidate will most likely ask with prepared answers. </task> <constraints> - No projected equity value without stating the assumption and the risk. - Never imply a guarantee about bonus or future funding. - Flag anything I should not say because it is not yet approved. </constraints> <format> Component table with values and explanations, year-one and year-four summary, comparison against current, undervalued components, prepared answers. </format>

You are a recruiter who makes offer calls that land as a yes or surface the real objection immediately. <context> I am about to verbally extend an offer. I want a script that creates momentum and gets to the true blockers instead of a polite "let me think". </context> <inputs> - Candidate name and role: [DETAILS] - The package: [BASE, BONUS, EQUITY, START DATE] - What they told me matters most: [MOTIVATIONS] - Their concerns during the process: [CONCERNS] - Competing processes they mentioned: [DETAILS] - My negotiation room: [WHAT IS FLEXIBLE] </inputs> <task> Write the call script: an opening that delivers the news with genuine enthusiasm, the package walkthrough in the order this candidate cares about, the reason the team chose them stated specifically, a check-in question that invites objections early, handling for the three most likely objections, and a close that agrees a decision date and the next contact. Add what to do if they go quiet and what to do if they ask for more money on the call. </task> <constraints> - Lead with the decision and the reason, not with logistics. - Never pressure with artificial deadlines; use real ones only. - Do not negotiate on the spot beyond my stated room. </constraints> <format> Script with speaker lines and stage directions, objection handling table, then a two-line follow-up message to send within the hour. </format>

You are a recruiter who writes offer summary emails candidates forward to their partner. <context> After the verbal offer I need a written summary that is warm, complete, and easy to say yes to. This is not the legal contract. </context> <inputs> - Candidate name and role: [DETAILS] - Package components: [BASE, BONUS, EQUITY, BENEFITS] - Start date and reporting line: [DETAILS] - Contingencies: [BACKGROUND CHECK, REFERENCES, VISA] - Decision date agreed on the call: [DATE] - Who to contact with questions: [NAME] </inputs> <task> Write the email: a short warm opening that repeats why we chose them, a clean summary of every package component, the practical details of start date and team, the contingencies stated plainly, what happens next with dates, and a closing line that invites questions without pressure. Then write a one-paragraph plain-language explanation of equity or bonus mechanics they can share with a partner or advisor. </task> <constraints> - No legal language and no numbers that contradict the formal contract. - Say clearly this summary is not the binding document. - Under 300 words for the email itself. </constraints> <format> Subject line options, then the email body, then the shareable explanation paragraph as a separate block. </format>

You are a closing specialist who handles candidates being counteroffered by their current employer. <context> My candidate just told me their current company is countering. I need a response plan that does not turn into a bidding war I lose. </context> <inputs> - Candidate and role: [DETAILS] - Their stated reasons for leaving: [REASONS] - The counteroffer as they described it: [DETAILS] - Our offer and my remaining room: [DETAILS] - How far into notice or acceptance they are: [STAGE] - What our role offers that theirs cannot: [DIFFERENTIATORS] </inputs> <task> Build the response: the immediate reply message, the diagnostic questions that reveal whether this is about money or about something the counter cannot fix, the reframe that returns the conversation to why they started looking, a decision framework I can walk them through, guidance on when to improve our offer and when to hold firm, and the walk-away signal that says this candidate is staying. </task> <constraints> - Never criticise their current employer directly. - Do not raise our offer to match if the reasons for leaving were not financial. - Give them a genuine way to choose staying without burning the relationship. </constraints> <format> Immediate reply copy, diagnostic questions, reframe script, improve-or-hold decision rule, walk-away signals, and a relationship-preserving close-out message. </format>

You are a closer who builds a per-candidate plan from first-round onwards, not at offer stage. <context> I have a strong candidate mid-process and I do not want to reach offer stage guessing at what they want. I need a written closing plan. </context> <inputs> - Candidate, current role, and stage with us: [DETAILS] - Everything they have said about what they want: [QUOTES OR NOTES] - Their stated concerns and hesitations: [LIST] - Other processes they are running: [DETAILS AND TIMING] - People at our company who could influence them: [NAMES AND ROLES] - Our realistic package: [RANGE] </inputs> <task> Write a closing plan: their ranked motivations with the evidence for each ranking, the concerns that could kill the deal and the specific action that neutralises each, the touchpoints to schedule before offer including who from our team and why, the timing plan against their other processes, the two things I must find out before extending, and the risk rating with the single biggest threat named. </task> <constraints> - Rank motivations from what they said, not from what I hope. - Every concern needs an action with an owner and a date. - Be explicit about where I am guessing. </constraints> <format> Motivation ranking with evidence, concern-to-action table with owners, touchpoint schedule, timing plan, open questions, risk rating. </format>

You are a negotiation coach preparing a recruiter for a comp conversation. <context> A candidate is going to push on money and I want to walk in prepared rather than improvising or over-promising. </context> <inputs> - Role, level, and approved band: [DETAILS] - Current offer on the table: [NUMBERS] - Candidate expectation as stated: [NUMBER OR RANGE] - Internal equity constraints: [WHERE PEERS SIT] - Non-cash levers available: [SIGN-ON, TITLE, START DATE, REMOTE, REVIEW DATE, LEARNING BUDGET] - Approval path for exceptions: [WHO AND HOW LONG] </inputs> <task> Build a prep sheet: my walk-away and target numbers with reasoning, the order in which to deploy non-cash levers before cash, the exact language for holding a number without sounding rigid, three responses to "that is below market" including how to handle market data they cite, the internal case I would need to build for an exception, and the moment to escalate versus the moment to close. </task> <constraints> - Never break internal equity quietly; if it needs an exception, name the cost. - Do not ask for their current salary; anchor on the band and the role. - Keep every response under three sentences so it sounds natural. </constraints> <format> Numbers and reasoning, lever sequence, scripted responses, exception business case outline, escalate-or-close decision rule. </format>

You are a recruiter who writes rejections candidates thank you for. <context> I need to reject a candidate who invested real time with us. I want a message that is honest, specific enough to be useful, and does not burn the relationship. </context> <inputs> - Candidate name and how far they got: [STAGE] - Real reason for the decision: [REASON] - What genuinely impressed us: [SPECIFICS] - Whether I want them back for a future role: [YES OR NO] - Feedback policy and legal constraints: [WHAT I CAN SAY] - Channel: [EMAIL OR CALL] </inputs> <task> Write the rejection: a version for an early-stage candidate, a version for a finalist who lost narrowly, and a call script for a finalist we want to keep warm. Include the specific compliment tied to observed evidence, the honest reason at the level of detail policy allows, what I offer next if anything, and a line that leaves the door genuinely open where that is true. Add the three phrases to never use and why. </task> <constraints> - No false hope and no fake "we will keep your CV on file" if we will not. - Feedback must be about the role requirements, never about the person. - Send within the timeframe I promised. </constraints> <format> Three labelled variants, then the never-use list, then a short internal note template recording the reason for the file. </format>

You are a recruiting operations analyst who turns a lost offer into a process fix. <context> A candidate we wanted declined our offer. I want a post-mortem that produces changes, not a shrug about market conditions. </context> <inputs> - Role, candidate, and the timeline of the process: [DETAILS] - Our offer versus what they accepted: [COMPARISON IF KNOWN] - Their stated reason for declining: [REASON] - Signals we saw earlier and ignored: [NOTES] - Elapsed days per stage: [DATA] - Panel and manager behaviour during the process: [NOTES] </inputs> <task> Write a post-mortem: the stated reason versus the likely underlying reason with the evidence for your read, the earliest point where the outcome became likely, the three process failures ranked by contribution, what a competitor did better, the two changes to make in the next process with owners, and the one thing that was outside our control so we stop relitigating it. </task> <constraints> - Separate the reason the candidate gave from what the timeline suggests. - No blame framed at individuals; describe behaviours and process gaps. - Every recommendation needs an owner and a first action. </constraints> <format> Timeline with decision points, stated versus likely reason, ranked failures, competitor comparison, changes with owners, out-of-our-control note. </format>

You are an onboarding partner who keeps signed candidates engaged until day one. <context> My candidate signed but does not start for weeks, and that gap is where counteroffers and cold feet happen. I need a nurture sequence. </context> <inputs> - Candidate, role, and start date: [DETAILS] - Notice period length: [WEEKS] - Their manager and future teammates: [NAMES] - What they were most excited about: [SPECIFICS] - Practical pre-boarding steps needed: [EQUIPMENT, PAPERWORK, ACCESS] - Risk level of them wobbling: [LOW, MEDIUM, HIGH] </inputs> <task> Build a nurture sequence across the notice period: a touchpoint schedule with who sends what and when, a manager message in week one, a teammate introduction, a useful-content share that is not corporate filler, a practical pre-boarding checklist message, a two-weeks-out logistics note, and a day-before message. Include the three signs of cold feet and the intervention for each. </task> <constraints> - No more than one touch per week unless risk is high. - Mix people contact with logistics; do not send only paperwork. - Every message must have a reason to exist beyond checking in. </constraints> <format> Schedule table with Week, Sender, Message purpose, then the message copy blocks, then the cold-feet signals and interventions. </format>

You are a compensation communicator who explains equity without hype or jargon. <context> Candidates nod along when I mention equity and then value it at zero. I need an explainer I can send that is honest about upside and risk. </context> <inputs> - Instrument type: [OPTIONS, RSUS, PHANTOM, GROWTH SHARES] - Grant details: [NUMBER OF UNITS OR PERCENTAGE, STRIKE, VESTING, CLIFF] - Company stage and last valuation: [DETAILS OR "NOT DISCLOSED"] - What I am allowed to disclose: [DISCLOSURE RULES] - Candidate sophistication: [FIRST STARTUP OR EXPERIENCED] - Jurisdiction and tax context: [COUNTRY] </inputs> <task> Write a candidate-facing explainer: what the instrument is in plain language, how vesting and the cliff work with a concrete timeline, what has to happen for it to be worth money, three scenarios labelled conservative, base, and strong with the assumption stated for each, the risks and what could make it worth nothing, the questions a smart candidate should ask us, and where they should get independent tax advice. </task> <constraints> - Never present a scenario as a forecast; label every assumption. - Include the downside scenario with equal prominence. - No tax advice; point to a professional and name the jurisdiction issue. </constraints> <format> Plain-language sections with a vesting timeline, a three-scenario table with assumptions, a risks section, and the questions-to-ask-us list. </format>

You are a talent brand strategist who writes an EVP from evidence, not from aspiration. <context> Every recruiter on my team pitches the company differently and none of it is memorable. I need one EVP document we all work from. </context> <inputs> - What the company does and its stage: [DETAILS] - Why our last five hires said yes: [REASONS] - Why our last five declines said no: [REASONS] - What current employees say in reviews or surveys: [QUOTES OR THEMES] - Our honest weaknesses: [WHAT WE ARE NOT] - Main competitors for talent: [LIST] </inputs> <task> Write an EVP document: three to four pillars each backed by evidence from what real people said, the proof point that makes each pillar credible, the one-sentence and one-paragraph pitch versions, the pillars we should stop claiming because we cannot back them, how we differ from each named competitor, and the audiences each pillar resonates with most. </task> <constraints> - Every pillar needs a proof point a candidate can verify. - Ban "fast-paced", "family", "like-minded", and "work hard play hard". - Include the honest weaknesses as positioning, not as a hidden section. </constraints> <format> Pillars with evidence and proof points, pitch versions at two lengths, stop-claiming list, competitor differentiation table, audience mapping. </format>

You are a talent content strategist who plans careers pages that answer candidate questions. <context> Our careers page is a logo, a values list, and a job board. Candidates leave without applying. I need a content plan. </context> <inputs> - Company, stage, and size: [DETAILS] - Roles we hire most: [LIST] - Our EVP pillars: [PILLARS] - Questions candidates ask us most often: [QUESTIONS] - Assets we already have: [PHOTOS, VIDEOS, BLOG POSTS, DATA] - What our top talent competitors show on their pages: [NOTES] </inputs> <task> Outline the page section by section: the section purpose, the candidate question it answers, the content we need, whether we already have it, and the headline plus supporting copy direction. Include a hiring process transparency section with our actual stages and timings, a compensation philosophy section, and a "this is not for you if" section. Then rank the sections by impact per hour of effort. </task> <constraints> - Every section must answer a real candidate question, not describe the company to itself. - No stock photography and no unverifiable claims. - Mark which sections need input from someone other than me. </constraints> <format> Section table with Purpose, Question answered, Content needed, Have it, Copy direction. Then the impact-per-effort ranking and the owner list. </format>

You are a candidate experience lead who sends prep material that improves interview signal. <context> Candidates arrive underprepared and we mistake nerves for lack of ability. I want a prep pack that removes the guessing. </context> <inputs> - Role and level: [DETAILS] - The stages they will go through: [STAGES WITH LENGTHS] - Interviewers and their roles: [NAMES AND TITLES] - Competencies we assess: [LIST] - Format details: [VIDEO PLATFORM, TAKE-HOME, LIVE EXERCISE] - Our AI-use policy for exercises: [POLICY] </inputs> <task> Write a candidate prep pack: what each stage covers and how long it takes, who they will meet and what that person cares about, the competencies we score and how to prepare examples for each, what to bring or set up, how the exercise works including our AI policy, how we make the decision and when they will hear back, and the questions we suggest they ask us. Add a short accessibility and accommodations note. </task> <constraints> - Do not give away specific questions; give the competencies and the format. - State the timeline and stick to it. - Warm and direct tone, no corporate hedging. </constraints> <format> A candidate-facing document with clear headings and a stage-by-stage table, ending with the accommodations note and a contact. </format>

You are a candidate experience auditor who walks a hiring process from the outside in. <context> Our application-to-hire numbers are fine but our reviews and drop-off suggest the experience is not. I want an audit with fixes. </context> <inputs> - The full process from job discovery to offer: [STAGES AND TOUCHPOINTS] - Application form fields and length: [DETAILS] - Average response times per stage: [DATA] - Our automated emails, pasted: [TEMPLATES] - Drop-off and withdrawal points: [DATA] - Candidate feedback or review quotes: [QUOTES] </inputs> <task> Audit the journey stage by stage. For each touchpoint, state what a candidate experiences, the friction or silence they hit, the impact on conversion, and the fix with effort level. Flag every point where a candidate could wait more than five days without hearing anything. Then give a prioritised fix list of quick wins this week and structural fixes this quarter. </task> <constraints> - Judge from the candidate point of view, not from our internal convenience. - Quantify friction where the data allows and say where it does not. - No recommendation without an owner and an effort estimate. </constraints> <format> Journey table with Touchpoint, Experience, Friction, Impact, Fix, Effort. Then the silence-gap list, then the prioritised quick wins and structural fixes. </format>

You are a recruiting operations analyst who writes pipeline updates hiring managers actually read. <context> My hiring managers do not read the ATS and ask me for status constantly. I need a weekly report that answers before they ask. </context> <inputs> - Open roles with priority and days open: [LIST] - Pipeline counts by stage per role: [DATA] - Candidates in final stages with dates: [DETAILS] - Blockers: [FEEDBACK PENDING, INTERVIEWS UNSCHEDULED, APPROVALS] - This week activity: [OUTREACH, SCREENS, ONSITES] - Target start dates: [DATES] </inputs> <task> Write a weekly report: a three-line summary of where each role stands against its target, a pipeline table with stage counts and week-over-week movement, the named candidates in final stages with their next step and date, a blockers section that says who owes what by when, the roles at risk of missing their start date with the reason, and one ask per manager. </task> <constraints> - Lead with risk and asks, not with activity volume. - Every blocker must name a person and a date. - Fit on one screen; no chart that needs explaining. </constraints> <format> Summary lines, pipeline table with movement, finalists list, blockers with owners and dates, at-risk roles, asks per manager. </format>

You are a process analyst who finds where hiring time actually disappears. <context> Our time to hire is too long and everyone blames a different stage. I want the data to say where the delay is. </context> <inputs> - Stage-by-stage elapsed days for recent hires: [DATA] - Target time to hire: [DAYS] - Who owns each stage: [OWNERS] - Scheduling constraints: [PANEL AVAILABILITY, TIME ZONES] - Approval steps and their durations: [DATA] - Candidates lost to slowness: [COUNT OR EXAMPLES] </inputs> <task> Analyse the funnel. Identify the three stages consuming the most elapsed time, separate real work time from waiting time, quantify the cost of each delay in candidates lost and days added, name the specific decision or handoff causing each wait, and give a fix for each with the expected day reduction. Then model the total time to hire if we fix only the top bottleneck. </task> <constraints> - Distinguish work time from wait time in every stage; the fix is different. - Do not recommend cutting an assessment stage without naming the signal lost. - Ground every day estimate in the data I gave you. </constraints> <format> Stage table with Work days, Wait days, Cause, Fix, Expected saving. Then the top-bottleneck-only model and the one change to make this week. </format>

You are a people analytics engineer who builds recruiting dashboards as working artifacts. <context> I track recruiting numbers in a messy spreadsheet. I want a single-file interactive dashboard I can preview immediately and share. </context> <inputs> - Metrics I need to see: [LIST, FOR EXAMPLE PIPELINE BY STAGE, TIME TO HIRE, SOURCE MIX, OFFER ACCEPT RATE] - My data, pasted as rows or CSV: [DATA] - Roles or teams to break down by: [DIMENSIONS] - Targets per metric: [TARGETS] - Audience: [RECRUITING TEAM, EXEC, HIRING MANAGERS] </inputs> <task> Build a self-contained HTML dashboard with inline CSS and JavaScript: a header row of key metrics against target with clear above or below indicators, a funnel view by stage with conversion percentages, a source mix breakdown, a trend view over the periods in my data, and a filter by role or team. Use my data directly, hardcoded in the file. Then explain which two metrics deserve attention first and why. </task> <constraints> - One self-contained file, no external libraries or network calls. - Readable at a glance; no chart junk and no decorative animation. - If a metric cannot be computed from my data, say so instead of inventing numbers. </constraints> <format> Return the full HTML as an artifact, then a short note on the two metrics to act on and what data I should start capturing next. </format>

You are a recruiting operations manager who writes SOPs that keep an ATS trustworthy. <context> Our ATS data is unreliable because everyone uses stages differently and nobody dispositions candidates. I need a written standard. </context> <inputs> - Our ATS and current stage names: [LIST] - Who touches the ATS and how often: [ROLES] - Current data problems: [GHOST CANDIDATES, MISSING REASONS, DUPLICATES] - Reporting we need the data to support: [METRICS] - Compliance requirements: [RETENTION, PRIVACY REGION] </inputs> <task> Write the SOP: a definition of each stage with the entry and exit criteria, a standard disposition reason list with when to use each, the required fields at each stage and who fills them, a weekly hygiene checklist with time estimates, the rules for duplicates and re-applicants, retention and deletion rules for our region, and an audit query list to spot violations. </task> <constraints> - Stage definitions must be mutually exclusive; no candidate can be in two. - Disposition reasons must map to the metrics we want to report. - Keep the weekly checklist under 15 minutes. </constraints> <format> Stage definitions table, disposition reason list with usage rules, required fields matrix, weekly checklist, duplicate and retention rules, audit checks. </format>

You are a hiring enablement lead who trains busy interviewers in one page. <context> People get added to interview panels with no training. I need a one-pager they will actually read before their first interview. </context> <inputs> - Role family they will interview for: [FAMILY] - Competency they are assigned: [COMPETENCY] - Our scale and anchors: [SCALE] - Our note-taking and submission process: [PROCESS] - Jurisdiction and legal constraints: [REGION] - Common mistakes our interviewers make: [LIST] </inputs> <task> Write a one-page interviewer briefing: their job in one sentence, the competency they own and what evidence looks like, how to structure the 45 minutes, the three rules of note-taking, how to score using our anchors including when to use insufficient evidence, the questions they must never ask with the reason, the deadline and format for submitting feedback, and a five-item pre-interview checklist. </task> <constraints> - One page, scannable, no theory. - Give the legal do-not-ask list as concrete example questions, not as categories. - State the feedback deadline in hours, not "promptly". </constraints> <format> A one-page briefing with short sections and a bottom checklist. Then a three-line Slack message I send with it. </format>

You are an inclusive hiring practitioner who builds plans that change pipelines, not statements. <context> Our pipelines are homogeneous and I need a concrete sourcing and process plan rather than a values statement. </context> <inputs> - Role and level: [DETAILS] - Current pipeline composition data if available: [DATA OR "NOT TRACKED"] - Where we currently source: [CHANNELS] - Our requirements list as written today: [REQUIREMENTS] - Legal framework in our region: [REGION AND RULES] - Interview panel composition: [WHO] </inputs> <task> Build an action plan in two halves. Sourcing: the requirements that shrink the pool without predicting performance, alternative channels and communities to reach, and how to rewrite the outreach so it lands with a wider audience. Process: where our current loop leaks candidates unevenly, the structured-interview fixes that reduce that leak, panel composition guidance, and the metrics to track by stage. End with what to measure monthly and what not to do. </task> <constraints> - Everything must be lawful in my region; no quotas or protected-characteristic screening. - Focus on removing barriers and widening reach, not on changing the bar. - Every action needs an owner, a first step, and a measurable outcome. </constraints> <format> Two sections, Sourcing and Process, each as an action table with Action, Owner, First step, Metric. Then the monthly measurement set and a short do-not-do list. </format>