Indeed

Job marketplace whose employer-side AI (the Talent Scout agent, Sourcing Assistant, Smart Screening and candidate matching) sources, ranks and scores candidates from Indeed's profile pool and applicant flow and drafts outreach to them.
SourcingScreening

HireAIScore rates Indeed 50 out of 100 (grade F, Substantial gaps) under rubric v1.0, last reviewed September 25, 2026. Indeed ranks 25 of 86 vendors rated in Candidate sourcing AI, against a category average of 46. That score is drawn from 40 evidence items across 7 rubric criteria for Indeed — 36 with a cited source, 5 recording that nothing was located.

Score breakdown

Indeed, criterion by criterion

Each rubric criterion is scored out of 100 on the cited evidence, then weighted into the total.

Weighted contribution shown to the right of each bar.

  1. Article 11 Technical Documentation

    Weight 20% · avg 48

    58

    +11.6

  2. Bias Audit Transparency

    Weight 18% · avg 40

    38

    +6.8

  3. FRIA Support

    Weight 15% · avg 31

    38

    +5.7

  4. Data Governance Disclosure

    Weight 15% · avg 53

    58

    +8.7

  5. Human Oversight Design

    Weight 12% · avg 56

    63

    +7.6

  6. Post-Market Monitoring

    Weight 12% · avg 40

    40

    +4.8

  7. Customer Documentation

    Weight 8% · avg 57

    64

    +5.1

Total weighted score

50.3(rounded to 50 for display)

Category avg is the mean raw score on that criterion across the 86 Candidate sourcing AI vendors in scope of this rubric, this one included.

Strongest criterion

Customer Documentation

64raw / 100

Contributes 5.1 to the total. 64 against a 57 category average.

Weakest criterion

FRIA Support

38raw / 100

Contributes 5.7 to the total. 38 against a 31 category average.

Cited evidence

The evidence behind the score

Every score is backed by at least one cited piece of evidence. Open a criterion to see what was found, when, and where.

Download diligence record

Evidence ledger

Items
40
Documentation
28
Public statement
6
Product feature
1
Absence
5
With a source URL
36 of 40
Source hostnames
7
Fewest items
4
FRIA Support

These figures measure how thoroughly Indeed was reviewed, not how Indeed performed — an absence row, recording that nothing was located, is counted like any other item.

Article 11 Technical Documentation

6 items58
  • Public statementCaptured Sep 25, 2026

    Indeed's Responsible AI page sets out five principles and says a cross-functional Responsible AI team tests products for 'fairness, safety and helpfulness before, during and after launch' and has engaged with NIST's AI Risk Management Framework.

    https://www.indeed.com/esg/responsible-ai

  • Public statementCaptured Sep 25, 2026

    Recruit Holdings publishes 'Indeed's Principles on Algorithms and Fairness', committing to approach fairness 'with rigor and care' through quantitative measurement as a company-wide effort, but gives no technical detail.

    https://recruit-holdings.com/en/sustainability/human-rights/policy/hr-technology/

  • DocumentationCaptured Sep 25, 2026

    A February 2026 explainer describes how the 0-100 Smart Fit Score turns employer criteria into a weighted framework (must-haves outweigh nice-to-haves), uses only applicant-supplied information, and does not rely on third-party data or comparisons with other candidates.

    https://www.indeed.com/news/releases/introducing-indeeds-smart-fit-score

  • DocumentationCaptured Sep 25, 2026

    The November 2025 'Job Matching 101' explainer outlines the matching engine's inputs (skills, experience, preferences and on-site behavior; 10,000+ attributes across 900 occupations), its three stages, and how it learns from outcomes such as employers moving candidates to interview.

    https://www.indeed.com/news/releases/how-indeed-job-matching-works

  • DocumentationCaptured Sep 25, 2026

    Indeed's SafeBase Trust Center lists an 'AI White Paper' on its lifecycle approach to AI safety, fairness and compliance, but the paper is gated behind a 'Get access' request rather than published.

    https://trustcenter.indeed.com/

  • AbsenceCaptured Sep 25, 2026

    No model or system cards, accuracy or subgroup performance metrics, EU-style instructions for use, or ISO/IEC 42001 certification were found on indeed.com's product, legal and ESG pages, in the Trust Center's compliance list, or on Recruit Holdings' sustainability pages.

    Have a URL that disputes this?

Bias Audit Transparency

5 items38
  • AbsenceCaptured Sep 25, 2026

    The ACLU/NYCLU crowd-sourced tracker of NYC Local Law 144 bias audits, updated in September 2026, contains no entry for Indeed, Glassdoor or Recruit Holdings.

    https://github.com/aclu-national/tracking-ll144-bias-audits

    Have a URL that disputes this?

  • Public statementCaptured Sep 25, 2026

    Indeed's AI/AEDT FAQ (updated 25 August 2026) says it does not consider any NYC offering an AEDT and that its Legal and Responsible AI teams 'regularly conduct assessments' of AI outputs across demographic groups, but it names no auditor and publishes no results.

    https://www.indeed.com/legal/http-indeed-com-legal-aiandaedtfaq

  • Public statementCaptured Sep 25, 2026

    Indeed's Smart Screening guide says the models 'are trained and audited for bias' but gives no auditor, date, methodology or impact ratios.

    https://www.indeed.com/hire/resources/howtohub/what-is-indeed-smart-screening

  • DocumentationCaptured Sep 25, 2026

    Indeed's privacy policy says voluntary demographic data (age, race and ethnicity, gender, disability and more) are collected with consent 'for evaluating our algorithmic fairness' and are not used in any determination of a job application.

    https://hrtechprivacy.com/brands/indeed

  • AbsenceCaptured Sep 25, 2026

    No third-party fairness audit (BABL AI, DCI, Holistic AI, Warden AI or similar) was found: trust.warden-ai.com/indeed returns 404, the Trust Center lists no bias-audit document, and searches of indeed.com and the web surfaced none.

    Have a URL that disputes this?

FRIA Support

4 items38
  • DocumentationCaptured Sep 25, 2026

    The FAQ answers how Indeed complies with the EU AI Act, the Colorado and Texas AI acts and the Illinois Human Rights Act only by citing a worldwide Prohibited Practices Policy and telling employers to review their obligations with counsel; it offers no FRIA or deployer template.

    https://www.indeed.com/legal/http-indeed-com-legal-aiandaedtfaq

  • DocumentationCaptured Sep 25, 2026

    The Smart Screening terms tell employers to keep criteria job-related and consistent with business necessity, avoid screening out people with disabilities, and treat scores as only one of several factors with 'meaningful human involvement', which is US anti-discrimination guidance for deployers rather than EU FRIA support.

    https://www.indeed.com/legal#SmartScreening

  • DocumentationCaptured Sep 25, 2026

    Indeed sells Smart Sourcing's AI candidate matching to German employers (EUR 400/month) on a page with no EU AI Act or deployer-obligation information.

    https://de.indeed.com/arbeitgeber/smart-sourcing

  • AbsenceCaptured Sep 25, 2026

    No Article 27 FRIA template, Article 26 deployer instructions or EU AI Act conformity material was found on indeed.com's US and German employer pages, in the Terms of Service or the AEDT FAQ, or among the Trust Center's public documents.

    Have a URL that disputes this?

Data Governance Disclosure

6 items58
  • DocumentationCaptured Sep 25, 2026

    The Trust Center lists SOC 2 Type 2, SOC 3, ISO/IEC 27001:2022, 27017, 27018 and 27701, CSA STAR Level 1 and the EU-US DPF, and publishes a subprocessor list naming OpenAI, but the SOC 2 and ISO 27001 reports require an access request and ISO/IEC 42001 is not listed.

    https://trustcenter.indeed.com/

  • DocumentationCaptured Sep 25, 2026

    The privacy policy (updated 22 September 2026) says personal data are used to train AI models, explains that Smart Screening's match score factors in text, audio or video follow-up answers, and gives users rights to contest automated decisions or opt out of profiling.

    https://hrtechprivacy.com/brands/indeed

  • DocumentationCaptured Sep 25, 2026

    Indeed states that protected-class signals such as age, gender and race, and third-party data, are not used in the Smart Fit Score calculation.

    https://www.indeed.com/hire/resources/howtohub/what-is-indeed-smart-screening

  • Public statementCaptured Sep 25, 2026

    Indeed describes a worldwide Prohibited Practices Policy that bars discrimination on protected characteristics, 'including using zip codes as proxies for protected classes', as well as biometric identification without consent and prediction of criminal offending.

    https://www.indeed.com/legal/http-indeed-com-legal-aiandaedtfaq

  • DocumentationCaptured Sep 25, 2026

    The Terms of Service publish a controller-to-processor Data Processing Addendum with EU SCCs, while the Talent Scout and Smart Screening terms let Indeed use Client Personal Data to develop, train and fine-tune its AI models.

    https://www.indeed.com/legal#dpa

  • DocumentationCaptured Sep 25, 2026

    Indeed's AI guide says matched candidates are generated from job descriptions, resumes, profiles and job-seeker activity on Indeed, and that generative features use OpenAI GPT models alongside Indeed's proprietary AI.

    https://www.indeed.com/hire/resources/howtohub/how-indeed-uses-ai

Human Oversight Design

7 items63

Post-Market Monitoring

5 items40
  • DocumentationCaptured Sep 25, 2026

    Public monthly Smart Screening release notes (April-August 2026) record changes to rule logic, automatic rescoring, ATS-delivery settings, fraud flags and new market pilots, but not model versions or fairness metrics.

    https://www.indeed.com/employers/cs/smart-screening/release-notes

  • DocumentationCaptured Sep 25, 2026

    A Vulnerability Disclosure Program web form and a Bugcrowd bug bounty give a public channel for security reports, but neither covers AI behavior or model incidents.

    https://www.indeed.com/security/vulnerability-disclosure-policy

  • Public statementCaptured Sep 25, 2026

    Indeed says it 'continually evaluates its AI models for performance and accuracy' and lets job seekers and employers give feedback on AI performance, but publishes no monitoring results.

    https://www.indeed.com/legal/http-indeed-com-legal-aiandaedtfaq

  • DocumentationCaptured Sep 25, 2026

    The privacy policy gives individuals a privacy-team email address for contesting any decision generated by automated processing and for opting out of profiling that has legal or similarly significant effects.

    https://hrtechprivacy.com/brands/indeed

  • AbsenceCaptured Sep 25, 2026

    No working public status page was found: status.indeed.com redirects to indeedstatus.com, which did not resolve in DNS on 2026-09-25, indeed.statuspage.io does not exist, glassdoor.statuspage.io is inactive, and no AI incident channel or model-update changelog was located.

    Have a URL that disputes this?

Customer Documentation

7 items64

Editorial notes

Company overview

Indeed, Inc. (founded 2004, headquartered in Austin, Texas, and a subsidiary of Recruit Holdings) runs the world's largest job site: 645 million job-seeker profiles and 3.3 million employers in more than 60 countries. This review covers its employer-side AI only. That includes the matching engine behind Matched Candidates and Advanced Matching, which decides which candidates employers see. Talent Scout, a conversational hiring agent announced in September 2025 and still in beta, ranks applicants, surfaces matches with explanations from 370 million sourceable profiles, and drafts outreach; it is US-only and runs inside Indeed, Workable and isolved, with Workday listed as coming soon. Sourcing Assistant (June 2026) finds and invites candidates around the clock under Smart Sourcing Professional ($520/month) or Enterprise plans. Smart Screening gives each application a 0-100 Smart Fit Score, an AI summary and optional AI Recruiter follow-up questions by text, audio or video. It is US-based, with pilots in Canada, the UK and Australia. Job-seeker tools such as Career Scout are out of scope.

Regulatory exposure

Talent Scout 'can help you rank applicants', and Smart Screening scores every application from 0 to 100. Employers can also have candidates who miss 'required' qualifications rejected automatically. These are the functions that NYC Local Law 144, California's FEHA automated-decision regulations and Illinois' amended Human Rights Act (in force since January 2026) regulate. Indeed's AEDT FAQ (updated 25 August 2026) says none of its NYC offerings is an AEDT. Its terms also bar employers from using Talent Scout as 'the sole or most important or substantial factor'. That wording tracks LL 144's 'substantially assist' test and leaves audit and notice duties with employers. AI Recruiter video answers feed the score, which could bring Illinois' AI Video Interview Act into play for Illinois roles. In the EU, Indeed sells Smart Sourcing's AI matching in Germany, and it says 70% of sponsored applications come from AI job recommendations. Both are uses (evaluating candidates and targeting job ads) that Annex III point 4 lists as high-risk; after the AI Omnibus, those obligations apply from 2 December 2027. Indeed would likely be the provider, yet it publishes nothing beyond a Prohibited Practices Policy. Colorado's replacement law, SB 26-189 (effective 1 January 2027), will require developers to give deployers technical documentation, and Indeed's nearest document, an AI White Paper, is gated.

Path to a higher score

The biggest gain would come from an independent bias audit of the Smart Fit Score, Talent Scout ranking and Matched Candidates, with impact ratios by sex, race/ethnicity and intersectional groups published in LL 144 format. Indeed already collects consented demographic data 'for evaluating our algorithmic fairness', so the data exist. It should also publish the gated AI White Paper alongside system cards for the matching, scoring and AI Recruiter models, covering training data, intended use, limitations, subgroup performance and update cadence. EU Smart Sourcing customers need AI Act instructions for use and a FRIA/DPIA template before December 2027. Indeed should restore a working public status page (status.indeed.com currently redirects to a domain that does not resolve) and add a model-change log and an AI incident channel. Finally, it should tell candidates that their AI Recruiter answers are scored.

Regulatory frame

What applies to candidate sourcing AI.

Ranking outputs are inside Annex III §4 when they influence hiring decisions. GDPR Article 22 (automated decision-making) and the GDPR rights around profiling apply independently for EU candidates.

Keep reading

All sourcing vendorsHow this score is builtAnnex IV section-by-section: what AI hiring vendors must documentWhat makes a defensible NYC LL 144 bias audit

Compare

Indeed against its nearest-scoring peers.

The closest scores in Candidate sourcing AI, side by side.

Build any comparison

Ranking

Others rated in Candidate sourcing AI

Ranked 25 of 86 by weighted total under rubric v1.0. The ordering is arithmetic on the rubric and carries no view on which tool suits a given hiring process.

  1. #22HiredScore52F
  2. #23Wellfound52F
  3. #24Textkernel51F
  4. #25IndeedThis profile50F
  5. #26Handshake50F
  6. #27Stepstone49F
All sourcing vendors

Conflicts of interest

No vendor pays for placement, scoring, or removal. Casework - the consulting firm that operates this directory - provides paid services to some vendors. Any active or recent (within 24 months) commercial relationship is disclosed on the affected vendor profile and the review is reassigned to an independent reviewer. See the full policy on About.

Casework has no commercial relationship with this vendor.