Your AI Action Plan

Create a research brief on [TOPIC] for [AUDIENCE]. Include: current landscape, 5 key facts, 3 risks, 3 open questions, and source links for every claim that affects a decision.

Review this research draft. Split it into: source-backed facts, reasonable inferences, unsupported claims, and questions to verify. Then rewrite the summary so unsupported claims are removed or clearly caveated.
Draft: [PASTE DRAFT]

Research [OPTIONS / VENDORS / TOPIC], compare them against [CRITERIA], and produce a recommendation memo. Include a table, tradeoffs, risks, source links, and the decision I should make if the priority is [COST / SPEED / QUALITY / RISK].

Create a research intake template for [ROLE / TEAM]. It should capture: decision to support, scope, time period, must-use sources, sources to avoid, comparison criteria, output format, approval owner, and caveats required before sharing.

Clean this data before I enter it into [SYSTEM]. Return a table with columns: [COLUMNS]. Standardize dates, names, categories, and phone/email formatting. Add a final column called Review needed for anything uncertain. Do not invent missing values.

Raw data:
[PASTE]

Map this incoming form to my database fields. Required destination fields: [FIELDS]. Validation rules: [RULES]. Return a field mapping, transformations needed, and exception cases that should stop for human review.

Extract records from this input and prepare them for [SYSTEM]. Return: 1. clean records table, 2. duplicate warnings, 3. missing required fields, 4. values that need human review, 5. a short change log.

Input:
[PASTE]

Create a data-entry QA playbook for [PROCESS]. Include: required fields, allowed formats, duplicate checks, sensitive-data warnings, examples of good records, examples of records to reject, and a final human review checklist.

Write a [TYPE - cold outreach / follow-up / proposal / status update] email.
From: [YOUR ROLE] at [COMPANY]
To: [RECIPIENT ROLE] at [THEIR COMPANY]
Context: [1-2 sentences of background]
Goal: [what you want them to do]
Tone: [professional / friendly / direct]
Length: [short = 3 sentences / medium = 1 short paragraph / full = structured email]

Write a weekly marketing performance report.
Period: [DATE RANGE]
Metrics to include: [list your KPIs]
Highlights: [what went well]
Issues: [what underperformed and brief reason]
Next week priorities: [3 bullet points]
Audience: [manager / team / client]
Tone: factual, no fluff. Use bullet points for metrics, short paragraphs for narrative.

Turn these rough notes into a clear [EMAIL / STATUS UPDATE / REPORT].
Audience: [WHO WILL READ IT]
Purpose: [DECISION, UPDATE, REQUEST, OR ESCALATION]
Raw notes:
[PASTE NOTES]

Return:
1. Suggested subject line
2. Short summary
3. Main message in my tone: [DIRECT / WARM / EXECUTIVE]
4. Action items with owners and dates
5. Risks or open questions

I'll share 3 examples of emails I've written. After reading them, identify:
1. My typical sentence length and structure
2. Words or phrases I use often
3. My tone (formal / casual / direct / warm)
4. Things I never say

[PASTE EMAIL 1]
[PASTE EMAIL 2]
[PASTE EMAIL 3]

Now write a [TYPE] email using my style. Here's the context: [CONTEXT]

Summarize this compliance-sensitive document for human review. Return: key facts, deadlines, obligations, missing information, ambiguous language, and questions for the responsible reviewer. Do not give legal, tax, or employment advice.

Document:
[PASTE]

Turn this policy or form instruction into a checklist. Include required fields, approvals, deadlines, evidence needed, common errors, and escalation triggers. Mark anything that needs a qualified human decision.

Text:
[PASTE]

Prepare a reviewer packet for this document. Return: 1. plain-English summary, 2. evidence table with source excerpts, 3. missing information, 4. risk questions, 5. reviewer decision log template. Keep all final decisions blank.

Document:
[PASTE]

Create a reusable compliance-review prompt for [ROLE]. It must require AI to: summarize only, cite source text, flag uncertainty, list missing information, avoid final legal/tax/employment decisions, and produce questions for the qualified reviewer.

Write a scheduling email for [MEETING PURPOSE].
Participants: [WHO]
Time options: [OPTIONS]
Duration: [LENGTH]
Prep needed: [PREP]
Tone: [FRIENDLY / DIRECT / FORMAL]
Include a 3-bullet agenda and a clear reply request.

Write a short message that shares my booking link and explains the purpose of the meeting.
Booking link: [LINK]
Purpose: [PURPOSE]
Who should attend: [ROLES]
What to prepare: [PREP]
Keep it under 120 words.

Create intake questions for a [MEETING TYPE] booking page. Then write a prep brief template that summarizes the answers into: context, goal, risks, decisions needed, and agenda.

Build a meeting routing playbook for [TEAM / ROLE]. Include: meeting types, when to book vs handle async, required attendees, intake questions, agenda templates, prep checklist, and follow-up owner.

N/A — manual.

You are helping a recruiter screen one candidate for a specific role. Do NOT decide whether to reject — score only.

Job description:
[PASTE JD HERE]

Rubric:
- MUST-HAVE qualifications: [LIST 3–5 items]
- NICE-TO-HAVE qualifications: [LIST 2–3 items]
- RED FLAGS / disqualifiers: [LIST 2–3 items]

Candidate resume:
[PASTE RESUME TEXT]

Output format:
1) Must-have matches (hit / miss with evidence quote)
2) Nice-to-have matches (hit / miss with evidence quote)
3) Red flags (if any)
4) Overall fit score 1–10 with one-line justification
5) Three follow-up questions I should ask this candidate

Do not recommend hire / no-hire. Flag anything that looks like protected-class information (age, family status, nationality, etc.) and note that I should remove it before any team discussion.

N/A — configuration-based. Set up the ATS AI feature, then in your weekly pipeline review use the following check:

For each AI-ranked shortlist I accepted this week:
- Does the shortlist skew on any protected attribute (age proxy, gender, etc.)?
- Did I override any AI ranking? Why?
- Any candidate that got a low score but a strong referral — was it re-reviewed?

Log these notes for the quarterly bias audit.

You are scoring a single anonymized resume against a predefined rubric.

Inputs (already masked):
- Redacted resume text
- Job ID and role type
- Rubric (must / nice / red flag)

Return JSON only:
{
  "must_hits": [{"criterion": "...", "evidence": "..."}],
  "nice_hits": [{"criterion": "...", "evidence": "..."}],
  "red_flags": [{"criterion": "...", "evidence": "..."}],
  "fit_score": 0,
  "confidence": "low|medium|high",
  "reviewer_questions": ["..."]
}

Do not output a hire recommendation. If the input still contains name, age, nationality, or other protected attributes after masking, set fit_score to null and flag "masking-failure".

N/A — manual.

You are helping a recruiter source for this role. Produce three things.

1) Three LinkedIn Boolean search strings — one tight, one broader, one that targets near-adjacent backgrounds. Explain who each one catches and misses.

2) Three outreach opener templates (under 90 words each) — one for exact-match candidates, one for career-switchers, one for senior folks who likely don't need the job. Each must:
- Reference a hook I can replace with something specific about the candidate's work.
- Name the role and one concrete reason it's interesting.
- End with a low-friction ask (quick 15-min intro call).
- Avoid clichés ("rockstar", "ninja", "we're changing the world").

3) A short checklist of red flags to avoid in the opener (spray-and-pray tone, salary bait, etc.).

Job description:
[PASTE JD HERE]

Target profile:
[PASTE NOTES ABOUT WHO YOU'RE LOOKING FOR]

Here is a shortlisted candidate profile and the role brief. Write one personalized outreach opener under 100 words.

Role brief:
[JD summary + why this role is interesting]

Candidate profile highlights:
- Current role: [TITLE at COMPANY]
- Recent public work / posts / projects: [BULLETS]
- Past signals that match the role: [BULLETS]

Constraints:
- One specific reference to their actual work in the first sentence.
- Name the role, company, and one concrete reason it's interesting.
- No generic praise ("impressive background").
- End with a 15-minute ask on a specific day/time window.

You are segmenting a candidate pool for outreach. Input:
- Target role and must-haves
- Talent Insights export rows (location, seniority, current employer, skills)
- Outreach volume budget this week

Return a JSON plan:
{
  "segments": [
    {
      "name": "...",
      "size": 0,
      "rationale": "...",
      "outreach_priority": "high|medium|low|skip",
      "opener_template": "...",
      "risk_notes": "..."
    }
  ],
  "do_not_contact": ["...current employer exclusions, recent candidates, etc."]
}

Never output real candidate names. Flag any segment where bias risk or legal risk (non-compete, protected class proxies) makes automated outreach unsafe.

N/A — manual.

You are writing a stage-change note for an ATS. Expand my shorthand into a structured rationale that is specific, neutral, and fair.

My shorthand: [SHORTHAND]
Role: [TITLE]
Stage change: [FROM_STAGE] → [TO_STAGE]

Output:
1) One-sentence summary (neutral tone).
2) Evidence bullets (behavior / answer / artifact).
3) Any follow-up next step.
4) Bias check: flag anything in my shorthand that references a protected attribute or a proxy for one, and suggest a rewrite.

Do not invent evidence I didn't mention.

N/A — configuration-based. Suggested weekly review:

1) Export this week's auto-rejected candidates from the ATS.
2) Spot-check 5 at random: was the reject reason accurate and fair?
3) Log any false-reject patterns for the next pipeline audit.

You are auditing a recruiter's weekly ATS export. Columns:
[candidate_id, role, stage, days_in_stage, last_action, recruiter, source, rejection_reason]

Return a short JSON report:
{
  "stale_candidates": [...top 10 with >14 days in stage],
  "inconsistent_reject_reasons": [...clusters of free-text reasons that should be merged],
  "bias_flags": [...any pattern worth a human look, e.g., one source dominating rejections],
  "suggested_actions": [...top 3 this week]
}

Do not include any PII beyond candidate_id.

N/A — manual.

You are writing a short, warm, specific candidate-facing message. Do NOT mention protected attributes. Do NOT fabricate feedback.

Type of message: [STATUS_UPDATE | REJECTION_AFTER_INTERVIEW | OFFER_EXTENSION | INTERVIEW_INVITATION]

Context:
- Role: [TITLE]
- Candidate's name and how we interacted: [NAME], [TOUCHPOINTS]
- Short internal notes to anchor feedback: [NOTES]
- Tone: professional, warm, human
- Constraints: under 150 words; no promises about future roles; no legal advice.

Output: subject line + email body + one-line internal note for the ATS record ("sent rejection after on-site, flagged strong communication—keep warm for later roles").

N/A — configuration-based. Suggested monthly check:

1) Pull the top 10 templates by send volume.
2) Re-read each for tone, clarity, and compliance with your hiring policy.
3) Re-run them through a quick Claude.ai review prompt: "Spot anything tone-deaf, biased, or legally risky in this template."

You are drafting a candidate-facing message triggered by an ATS stage change.

Input JSON:
{
  "candidate_first_name": "...",
  "role_title": "...",
  "from_stage": "...",
  "to_stage": "...",
  "recruiter_shorthand_note": "..."
}

Output JSON:
{
  "subject": "...",
  "body": "...",
  "needs_human_review_reason": "...",
  "risk_flags": ["bias", "legal", "tone"]
}

Always set needs_human_review_reason. Never produce a final send.

N/A — manual.

You are preparing onboarding communications. Do not change any legal terms. If you are not sure whether something is a legal term, flag it for HR review instead of rewriting.

Inputs:
- Offer summary: [TITLE, START DATE, COMP SUMMARY, LOCATION, MANAGER]
- New-hire preferences: [PRONOUNS, REMOTE/HYBRID, ANY ACCOMMODATIONS SHARED]
- Company-approved welcome-email template: [PASTE]

Output:
1) A personalized welcome email under 200 words.
2) A first-day agenda outline.
3) Plain-language summary of the 3 most confusing offer terms, with a "talk to HR if unclear" closer.
4) Flags: anything in the inputs that looks like a legal term I should not rewrite.

N/A — configuration-based. Quarterly audit checklist:

1) Pull a list of new hires onboarded this quarter.
2) Confirm I-9s completed within the federal 3-day window.
3) Confirm state-specific forms were triggered by the HRIS.
4) Spot-check 3 hires for document retention.

You are running a document-completeness check on a set of new-hire onboarding records (PII masked).

Input JSON schema:
[
  {
    "hire_id": "...",
    "start_date": "YYYY-MM-DD",
    "documents": [{"type": "I-9|W-4|state|handbook|direct_deposit", "status": "signed|pending|overdue|missing"}],
    "state": "US state code"
  }
]

Return a prioritized action list:
{
  "deadline_risks": [...records where I-9 or state form is at risk],
  "completeness_issues": [...missing doc combinations],
  "name_mismatch_flags": [...between offer, I-9, and payroll],
  "suggested_next_actions": ["..."]
}

Never output PII. Never claim a record is legally compliant — only flag for HR review.