Career

How to Write a Job Description with AI

Write an inclusive, compelling job description that attracts qualified candidates and accurately represents the role.

A weak job description either attracts too few applicants or attracts the wrong ones. AI can write role descriptions that clearly articulate responsibilities, set realistic expectations, use inclusive language that broadens the candidate pool, and position the company in a way that excites strong candidates who have options.

How job descriptions silently filter out great candidates

Most job descriptions are written by hiring managers under time pressure, which means they default to copying a previous JD or drafting a requirements list without thinking about how candidates will read it. The result is a list that conflates genuine requirements with preferences, uses credential inflation as a shortcut for quality signals (requiring a degree for roles where experience matters more), and includes gendered language patterns that research shows reduce applications from women by up to 30% even when the workplace is genuinely inclusive. The further problem is competitive positioning. A strong candidate — who has options — will read the JD and ask: does this role sound interesting? Does this company sound like a place where I would grow? Most JDs fail to answer either question. They describe the job as it exists today, not as it could develop, and they describe the company in three boilerplate sentences that sound like every other company.

How AI writes clearer, fairer requirements

AI's most useful role in JD writing is auditing requirements for necessity. When you paste a requirements list and ask 'which of these requirements could a strong candidate lack on day one but learn within 6 months?', it consistently identifies 2 to 4 items that are being treated as mandatory but are really preferences. Reclassifying these as preferred broadens the candidate pool without reducing quality — candidates who meet 8 of 10 requirements but are strong performers are often better hires than candidates who match all 10 but bring no growth potential. AI is also effective at flag-reviewing gendered phrasing. Words like 'rockstar', 'ninja', and 'aggressive growth targets' have measurably different effects on different candidate segments, and AI can suggest neutral alternatives instantly.

What inputs produce the best output

Before prompting, gather four things: a description of what success looks like in the first 90 days (not a list of responsibilities — an outcome), the actual must-have technical skills verified by the hiring manager, the team structure the person will join, and one honest sentence about why a great candidate would choose this role over a competitor's. That last input is the hardest and most valuable. If you cannot articulate why a high-performing person would accept this offer, that is a positioning problem AI cannot solve for you — but it can help you identify and articulate your actual differentiators once you are honest about what they are. With those four inputs, AI can generate a complete JD in a format that both ATS systems can parse and humans want to read.

Step-by-step guide

Define the role clearly

Specify the team, reporting structure, core responsibilities, and what success looks like in 90 days.

Write requirements carefully

Ask AI to separate genuine requirements from nice-to-haves and flag any requirements that may deter qualified candidates.

Apply inclusive language

Ask AI to audit the draft for gendered language, unnecessary credential inflation, and exclusionary phrasing.

Add a compelling company section

Ask AI to write a 3 to 4 sentence company pitch that explains why a strong candidate should choose you.

Ready-to-use prompts

Full job description from role brief

Act as a senior HR business partner. Write a job description for a [JOB TITLE] role at [COMPANY TYPE, e.g. Series A SaaS startup / enterprise financial services firm]. Team: [TEAM SIZE AND STRUCTURE]. Core responsibilities: [LIST 3-5 KEY RESPONSIBILITIES]. Must-have requirements: [GENUINE REQUIREMENTS ONLY]. Nice-to-have: [PREFERENCES]. Salary range: [RANGE]. What success looks like in 90 days: [OUTCOMES]. Why a strong candidate should choose this role: [DIFFERENTIATOR]. Format: role summary paragraph, 5 responsibilities, requirements split as must-have and nice-to-have, a company pitch of 3 sentences. Avoid: credential inflation, gendered language, buzzwords like 'rockstar' or 'ninja'.

Why it works

Separating must-have from nice-to-have in the input forces the hiring manager to be honest about requirements before writing, which produces a JD that attracts a broader pool without lowering the bar.

Inclusivity audit of existing JD

You are a diversity and inclusion hiring consultant. Audit this job description for: (1) gendered language patterns that research shows reduce applications from women or non-binary candidates, (2) requirements that are proxies for credentials rather than actual skills, (3) culture language that signals a narrow culture fit rather than value alignment, (4) anything that could deter candidates from underrepresented backgrounds. For each issue found: quote the specific phrase, explain why it is problematic, and provide a revised version. JD to audit: [PASTE JD].

Why it works

Asking for specific quotes and revised versions makes the feedback immediately actionable rather than general — the hiring manager can make targeted changes without rewriting the whole document.

Practical tips

✦Write 'what success looks like in 90 days' before writing the requirements list — this outcome-first thinking clarifies which requirements are real and which are just habits from the last JD.
✦Ask AI to split your requirements into genuine must-haves versus preferences, then move anything into nice-to-have that a strong hire could learn in 6 months on the job.
✦Generate 3 variants of the company pitch paragraph — one emphasizing mission, one emphasizing growth opportunity, one emphasizing team quality — and test which resonates most with your target candidate profile.
✦After writing the JD, ask AI to write the job from the perspective of a strong candidate who is currently employed and deciding whether to apply — this surfaces gaps in your positioning.
✦Ask AI to identify any phrase in the requirements that could function as a proxy filter for age, background, or institution rather than actual job performance.