👥 HR & Recruiting Prompt Generator Free · No Signup

AI Job Description Generator — From Intake Notes to a Clean, Inclusive JD

⚡ Quick Answer

An AI job description generator takes your intake notes — the real responsibilities and genuine must-haves — and drafts a structured job description with those facts only. This generator adds what listicles don't: a hard ban on invented requirements, a gender-neutral language pass, and a bias-flag report showing exactly what the AI changed and why.

⚡ AI Job Description Generator — From Intake Notes to a Clean, Inclusive JD

Configure your facts below to generate a production-ready, constraint-locked prompt.

100% Client-Side

What this generator does

You paste in the raw material from your hiring-manager intake — the bullet points, the half-finished notes, the "we need someone who can…" list — and the generator builds a complete prompt that turns those notes into a publication-ready job description. The output is structured the way strong job descriptions are structured: an honest overview, the real responsibilities, required qualifications separated cleanly from preferred ones, what success looks like in the first 90 days, and what the role offers.

The difference from a generic "write me a job description" prompt is the second half of the output: a changes and bias-flag report. The model must list every phrase it reworded for inclusivity, every claim it could not trace back to your intake, and every question it needs the hiring manager to answer. You see the draft and the audit trail together, so nothing the AI quietly added slips into a live posting.

Who it's for

How it works

  1. Start from intake, not from a title. A job title alone tells the model nothing true about your role. Paste the actual responsibilities and must-haves from your hiring conversation.
  2. Separate required from preferred honestly. If a qualification is a genuine deal-breaker, put it in Required. Everything else goes in Preferred. The template enforces the split.
  3. Add the 90-day success picture. "By day 90 this person will have shipped X and own Y" is the single most useful section in a job description — candidates self-select against it, and it costs you two sentences.
  4. Generate and copy the prompt. Run it in the recommended model. You get the job description, then the bias-flag report.
  5. Work the report. Answer the model's questions, strike anything it flagged that you can't verify, and only then publish.

The prompt template

PROMPT TEMPLATE 
You are an expert recruitment copywriter and inclusive-hiring specialist.

TASK: Write a complete job description using ONLY the information supplied below.

INPUTS:
- Job title: [job title]
- Department / team: [team]
- Location and work model: [location, onsite/hybrid/remote]
- Employment type: [full-time / part-time / contract]
- Key responsibilities (from hiring-manager intake): [paste raw notes]
- Required qualifications — genuine must-haves only: [list]
- Preferred qualifications: [list]
- Salary range: [range, or "not provided"]
- What success looks like in the first 90 days: [describe, or "not provided"]
- Tone: [professional / friendly / direct]

STRUCTURE (use these headings, in this order):
1. Headline options — give me 2 alternative job titles/headlines.
2. Overview — 3–4 sentences: what the team does, why this role exists, who it works with.
3. Key responsibilities — bullets rewritten from my intake notes, each starting with a verb.
4. Required qualifications — ONLY items from my Required list. Nothing else belongs here.
5. Preferred qualifications — ONLY items from my Preferred list.
6. What success looks like in 90 days — from my input; if I wrote "not provided", write "[TO CONFIRM WITH HIRING MANAGER]" and move on.
7. What we offer — ONLY benefits or facts I supplied. If I supplied none, write one neutral sentence and flag it in the report.

LENGTH: 450 words maximum for sections 2–7 combined. Short sentences. No paragraph over 4 lines.

INCLUSIVE LANGUAGE RULES:
- Gender-neutral throughout (they/them; no "he/she", no gendered job titles like "salesman").
- Replace coded language: no "rockstar", "ninja", "guru", "wizard", "high-energy", "young and energetic", "digital native", "manpower", "culture fit", "fast-paced" (say what the pace actually involves), "competitive" as a personality demand.
- Describe physical or sensory requirements ONLY if I listed them as essential functions, and phrase them as functions ("must be able to lift 15kg unaided as part of warehouse duties"), never as body or health demands.
- If my intake notes contain coded or gendered phrases, rewrite them neutrally AND list each one in the report.

NEGATIVE CONSTRAINTS — do not break these:
- Do NOT add any requirement, tool, degree, certification, licence, or years-of-experience figure that is not in my inputs. This is the most important rule in this prompt.
- Do NOT invent salary figures, benefits, team size, reporting lines, travel percentages, or company facts.
- Do NOT promise visa sponsorship, relocation, bonuses, or equity unless I stated them.
- Do NOT use superlatives about the company ("world-class", "industry-leading") unless I supplied them.

FINAL SECTION — "CHANGES & BIAS-FLAG REPORT":
1. Phrases you reworded for inclusivity (original → replacement).
2. Anything in my inputs that sounds like a requirement but is vague (e.g. "senior-level", "strong communicator") — quote it and ask what observable evidence would prove it.
3. Questions the hiring manager must answer before this is published.
Best model for this: Claude Sonnet 5.5 (long instruction-following and nuanced language review). GPT-6.1 is a strong alternative when you need many JDs in a consistent house format.

Worked example

Filled inputs: Job title: Senior Customer Success Manager. Team: Customer Success, B2B SaaS. Location: Bengaluru, hybrid (3 days in office). Responsibilities (raw notes): "own a book of ~40 mid-market accounts, run QBRs, spot churn risk early, work with support on escalations, onboard new accounts with solutions engineer". Required: "4+ years in customer success or account management for B2B software; experience running executive business reviews; comfort with usage data and spreadsheets". Preferred: "SaaS onboarding experience; Hindi or another Indian language". 90-day success: "Own the full book independently, complete first QBR cycle, build a churn-risk tracker for the book". Salary: ₹28–34 LPA.

Output excerpt (Overview + one flagged item):

Overview: Our Customer Success team partners with mid-market customers after the sale to make sure they get lasting value from the platform. As Senior Customer Success Manager you'll own a book of around 40 accounts — running quarterly business reviews, spotting churn risk early, and working with Support and Solutions Engineering so customers are onboarded well and stay.
>
Changes & bias-flag report (excerpt): 1. Your notes said "spot churn risk early" — kept as-is; consider adding the specific signal you use (usage drop, ticket volume) so candidates can self-assess. 2. No coded language found. 3. Question: your Required list says "4+ years" — is the year count itself essential, or would equivalent experience running QBRs for a similar book size qualify? Years-of-experience filters are the most common accidental pool-shrinker in job descriptions.

Guardrails

Frequently Asked Questions

Practical answers covering model choice, workflow integration, and compliance rules.

The facts that make this role different from every other posting with the same title: the team and what it does, the real responsibilities from your intake, the genuine must-haves kept separate from nice-to-haves, the location and work model, the salary range if you have one, and what success looks like in the first 90 days. The more of your input is fact rather than adjectives, the less the model has to invent.

Say so explicitly, and make the model prove compliance. This template carries the rule twice — once in the constraints ("do not add any requirement not in my inputs") and once in the output spec, where the model must list questions instead of filling gaps. The pattern generalises: a prohibition plus an audit section catches far more invention than a prohibition alone. If a model still adds a degree or a year count, delete it and reply "remove every requirement not in my original list" — it will comply.

The recurring offenders are coded rather than openly discriminatory: "rockstar", "ninja" and "guru" (gendered and off-putting), "high-energy", "young and energetic" and "digital native" (age-coded), "manpower" and gendered titles (gendered), and "culture fit" (an invitation to hire similarity). Requirements can bias too — a degree demanded for work that doesn't need one filters on socio-economic background. The generator's language pass flags these categories and shows you every change.

Yes, whenever you can. A stated range is the single clearest signal a posting gives a candidate, several jurisdictions now require pay transparency in job postings, and a range also disciplines the AI — it stops the model writing "competitive salary" filler. If your organisation won't publish a range, leave the field empty and the template writes one neutral sentence and flags it in the report rather than inventing a number.

Yes — paste the existing posting into the responsibilities field and say so in the title line. The template treats it as source material: it keeps every traceable requirement, drops or flags anything it cannot verify, restructures into the standard headings, and the bias-flag report shows you what the original got wrong. This is the fastest way to clean up a legacy JD library.

Internal links

---