Job Description Bias And Risk Auditor

Review job descriptions for observable bias, pay transparency, location clarity, and compliance risk signals before they are published.

This Actor helps recruiting teams, HR operators, job boards, agencies, and compliance reviewers turn draft job descriptions into structured review findings. It highlights potentially exclusionary language, vague compensation wording, unclear remote or location requirements, risky legal phrasing, and missing job basics. Each result includes evidence snippets, severity, rewrite suggestions, and recommended next steps for human review.

It does not certify legal compliance. It surfaces wording and disclosure signals that may deserve HR, legal, DEI, or recruiting-ops review.

What It Checks

• Bias and exclusionary wording, including age-coded phrases, native-speaker requirements, culture-fit phrasing, and informal coded labels such as "rockstar" or "ninja"
• Pay transparency signals, including missing salary ranges and vague compensation phrases such as "competitive salary"
• Location clarity, including missing remote, hybrid, onsite, timezone, travel, and eligibility details
• Legal-risk wording, including physical requirements, citizenship language, background-check wording, and unbounded work-hour expectations
• JD clarity, including missing responsibilities, requirements, employment type, or mismatched seniority and experience requirements

Inputs

Job Description

Paste the full job description into the large text area. The Actor returns one output row with findings, evidence, and rewrite suggestions.

Include the title, responsibilities, requirements, location, work mode, pay range, benefits, and employment type when available. Short or incomplete drafts can still be checked, but the Actor may warn that missing-field findings could reflect incomplete input.

Dataset ID

Optionally append results to an existing Apify dataset in addition to the default run dataset. This is useful for recurring recruiting audits, agency workflows, or job-board moderation pipelines.

Run ID

Optionally pass an upstream workflow, client, or batch ID. The Actor copies it to each output row.

Output

The dataset returns one row per job description run.

Example fields:

{
  "jobTitle": "Senior Growth Rockstar",
  "overallRiskScore": 100,
  "riskLevel": "critical",
  "biasFindings": [
    {
      "category": "bias",
      "severity": "high",
      "issue": "Age-coded language may discourage older candidates or imply age preference.",
      "evidence": "looking for a young and energetic digital native to join our fast-paced team",
      "suggestedRewrite": "Describe the skill or experience needed without age-coded wording..."
    }
  ],
  "payTransparencyFindings": [
    {
      "category": "payTransparency",
      "severity": "medium",
      "issue": "Compensation wording appears incomplete or vague.",
      "evidence": "Competitive salary",
      "suggestedRewrite": "Use a specific range, such as \"GBP 55,000-70,000 base salary plus bonus eligibility\"."
    }
  ],
  "locationFindings": [],
  "legalRiskFindings": [],
  "clarityFindings": [],
  "suggestedRewrites": [],
  "recommendedActions": [
    "Review high-severity findings before publishing or sending the job description to candidates."
  ],
  "evidenceSnippets": [],
  "warnings": [],
  "analysisStatus": "success",
  "runId": "weekly-review",
  "checkedAt": "2026-05-15T10:00:00.000Z"
}

Risk Score

The Actor assigns an observable risk score from 0 to 100.

• low: few or no visible review signals
• medium: wording or missing information should be reviewed before publishing
• high: several material findings or at least one high-severity issue
• critical: many findings across multiple categories, requiring human review before publishing

Scores are designed for triage. They are not legal determinations.

Common Use Cases

• Review job descriptions before they go live
• Audit a batch of job postings for pay transparency and wording risk
• Build a recruiting quality-control workflow in Apify
• Give agencies or HR teams evidence-backed rewrite suggestions
• Track recurring JD issues across roles, teams, or clients
• Moderate job-board submissions before publication

API Example

Run the Actor from another system with a normal Apify API call:

curl "https://api.apify.com/v2/acts/trovevault~jd-bias-compliance-risk-auditor/runs" \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "jobDescription": "Customer Success Manager\n\nWe are hiring a full-time Customer Success Manager in London. Salary range: GBP 45,000-55,000. You will onboard customers, manage renewals, and partner with product teams. Hybrid role, two office days per week.",
    "runId": "weekly-jd-review"
  }'

Daily Test Behavior

The default input contains a synthetic job description that intentionally includes several review signals. This lets Apify daily tests run without external credentials, paid APIs, or private HR data while still producing a non-empty dataset.

For real use, replace the sample with your own job descriptions.

Limitations

• The Actor reviews text supplied in the input. It does not fetch ATS records or scrape job boards in this version.
• Bias, clarity, and disclosure findings are global review signals. Local, sector, union, or role-specific requirements may still be stricter.
• It detects observable wording and missing disclosure signals. It cannot certify that a job description is legally compliant.
• Rewrite suggestions should be reviewed by the responsible HR, legal, DEI, or recruiting team before publication.
• Very short drafts can trigger missing-information findings because the Actor cannot know whether omitted sections exist elsewhere.

Tips For Better Results

• Paste the complete JD, not only the requirements section.
• Include compensation, location, work mode, travel, benefits, employment type, and work authorization language when available.
• Run the Actor before publication and again after material edits.
• Use runId to connect results to your ATS, job board, agency client, or internal review workflow.