Recruitly
HireAIScore rates Recruitly 41 out of 100 (grade F, Substantial gaps) under rubric v1.0, last reviewed September 25, 2026. Recruitly ranks 61 of 77 vendors rated in ATS-integrated AI, against a category average of 47. That score is drawn from 37 evidence items across 7 rubric criteria for Recruitly — 32 with a cited source, 5 recording that nothing was located.
Score breakdown
Recruitly, 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.
Article 11 Technical Documentation
Weight 20% · category avg 51
45
+9.0
Bias Audit Transparency
Weight 18% · category avg 41
15
+2.7
FRIA Support
Weight 15% · category avg 32
35
+5.3
Data Governance Disclosure
Weight 15% · category avg 54
45
+6.8
Human Oversight Design
Weight 12% · category avg 57
60
+7.2
Post-Market Monitoring
Weight 12% · category avg 42
45
+5.4
Customer Documentation
Weight 8% · category avg 59
58
+4.6
Total weighted score
40.9(rounded to 41 for display)
Category avg is the mean raw score on that criterion across the 77 ATS-integrated AI vendors in scope of this rubric, this one included.
Strongest criterion
Human Oversight Design
60raw / 100
Contributes 7.2 to the total. 60 against a 57 category average.
Weakest criterion
Bias Audit Transparency
15raw / 100
Contributes 2.7 to the total. 15 against a 41 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 recordEvidence ledger
- Items
- 37
- Documentation
- 25
- Public statement
- 5
- Product feature
- 2
- Absence
- 5
- With a source URL
- 32 of 37
- Source hostnames
- 4
- Fewest items
- 4
- Bias Audit Transparency
These figures measure how thoroughly Recruitly was reviewed, not how Recruitly performed — an absence row, recording that nothing was located, is counted like any other item.
Article 11 Technical Documentation
6 items45
- DocumentationCaptured Sep 25, 2026
The AI Application Review page describes Overall, Skills and Language scores with written explanations for each category. It names Qdrant, OpenAI and Anthropic as the technology. It claims 'over 90% agreement with experienced human recruiters' but publishes no method.
- DocumentationCaptured Sep 25, 2026
The AI Candidate Match page lists its ranking factors: skills, experience, culture fit indicators, career trajectory, salary alignment, location preferences and availability. It names Qdrant, OpenAI embeddings and Claude, and says the model keeps learning from successful placements, recruiter behaviour and candidate outcomes.
- DocumentationCaptured Sep 25, 2026
A public benchmarks page scores 20 language models on sixteen recruitment test cases, including a discriminatory-request refusal case and a candidate-to-job match rationale case. Scoring uses machine checks plus a blind model judge, and the page reports latency, cost and drift from a month of production traffic.
- DocumentationCaptured Sep 25, 2026
An engineering post from September 2026 says TypeSafe's Jev model now scores skills, experience, location, visa and deal breakers when a candidate is reviewed against a job, and also gives scorecard verdicts. Each decision carries a confidence score, has to pass a fixed evaluation battery, and falls back to the previous language-model path.
https://recruitly.io/resources/decision-model-in-production-the-day-it-was-released
- DocumentationCaptured Sep 25, 2026
The AI hub lists six AI capabilities and 12 agents. It gives no high-risk classification, conformity-assessment status or limitations statement, even though Recruitly's EU and UK regulatory guides send readers there for its 'position on the AI features'.
- AbsenceCaptured Sep 25, 2026
No model card, instructions for use, intended-purpose statement or ISO/IEC 42001 certification was found.
Where we looked: the recruitly.io sitemap (754 URLs), feature and security pages, llms.txt and llms-full.txt, and the sitemaps of help.recruitly.io and help.recruitly.app.
Bias Audit Transparency
4 items15
- AbsenceCaptured Sep 25, 2026
No NYC LL 144 bias audit, independent fairness audit or adverse-impact analysis of AI Application Review or AI Match scores was found.
Where we looked: recruitly.io and its sitemap, llms-full.txt, help.recruitly.io and help.recruitly.app. trust.warden-ai.com/recruitly returns 404.
- Public statementCaptured Sep 25, 2026
The product FAQ says AI 'excels at consistent, bias-free screening at scale' and eliminates 'many forms of unconscious bias', but publishes no test results or audit to support either claim.
- DocumentationCaptured Sep 25, 2026
The only published fairness-related test is TC8, which checks that a model refuses to write a discriminatory job advert. It measures prompt refusal, not score gaps between candidate groups.
- Public statementCaptured Sep 25, 2026
Recruitly's UK guide (25 May 2026) tells agencies that every AI tool should come with a written statement on bias testing, and points to Recruitly's AI product page for its position. That page contains no bias-testing statement.
FRIA Support
4 items35
- DocumentationCaptured Sep 25, 2026
Recruitly's EU AI Act guide (25 May 2026) lists deployer duties (follow the provider's instructions, real human oversight, candidate transparency, logs, monitoring) and a checklist for the quarter. It never mentions Article 27 impact assessments and wrongly dates recruitment AI obligations to 2 August 2027 under Article 6(1).
- DocumentationCaptured Sep 25, 2026
A multi-jurisdiction parent guide gives agencies five generic actions: inventory AI use, tell candidates, keep human oversight, audit for bias, and get vendor statements in writing. None is mapped to Recruitly's own features.
- DocumentationCaptured Sep 25, 2026
The security page says the platform has tools to carry out and document DPIAs and keep an Article 30 processing register. That is GDPR tooling, not an AI Act fundamental rights impact assessment template.
- AbsenceCaptured Sep 25, 2026
No FRIA template, Article 26/27 deployer pack or instructions for use was found on recruitly.io, help.recruitly.io or help.recruitly.app.
Data Governance Disclosure
5 items45
- DocumentationCaptured Sep 25, 2026
The security page names Google Cloud Belgium (europe-west1) as the primary data centre, Google London (europe-west2) as secondary, and MongoDB Cloud. It lists AES-256 at rest, TLS 1.3, and a 30-day retention period after deactivation or termination. It says the ISO 27001 and SOC 2 Type II certifications belong to GCP and MongoDB, not Recruitly.
- DocumentationCaptured Sep 25, 2026
The GDPR policy says Recruitly acts as data processor, that its infrastructure is hosted on Amazon Web Services with backups in two EU locations, and that breaches are reported to customers and the ICO within 72 hours.
- DocumentationCaptured Sep 25, 2026
The Terms make Recruitly a processor under a Data Processing Agreement that is referenced but not published. They bar customers from uploading special-category data such as racial or ethnic origin, health or religious beliefs, and say partial data may stay in backups for up to 50 days.
- Public statementCaptured Sep 25, 2026
A security explainer (updated 20 Sep 2026) says customer data is not used to train anyone's model. It says every AI call goes through a layer Recruitly controls, checked by a code scan that found more than forty AI touchpoints missing from the hand-kept list.
- AbsenceCaptured Sep 25, 2026
No sub-processor list naming AI model providers or routers was found, and recruitly.io/sub-processors, /subprocessors and /dpa all return 404. Feature pages name OpenAI, Anthropic and Qdrant, an engineering post names TypeSafe's Jev, and the benchmarks page shows models reached via OpenRouter, Google AI Studio and Telnyx.
Human Oversight Design
7 items60
- DocumentationCaptured Sep 25, 2026
The screening-policy help article (updated 8 Jul 2026) says an agency's written criteria are applied to every applicant but nobody is rejected automatically: 'the AI scores and explains; you decide'.
- DocumentationCaptured Sep 25, 2026
The AI review help article (updated 31 Aug 2026) shows a fit summary with strengths and gaps for each applicant. It says the AI never decides for the recruiter, who approves or rejects manually.
- DocumentationCaptured Sep 25, 2026
Match results show reasons (skills, experience, location) for each candidate scored against every live job, and each accept or reject decision by a recruiter is recorded.
- DocumentationCaptured Sep 25, 2026
The AI meeting assistant shows AI scores as drafts until a human saves them, attaches the transcript quote behind each score, and records changes as human overrides.
- DocumentationCaptured Sep 25, 2026
The copilot drafts emails, notes and updates but waits for a recruiter's confirmation before acting, and each confirmed action is logged, attributed to the person who confirmed it and can be undone.
https://recruitly.io/resources/why-our-ai-copilot-asks-before-it-acts
- DocumentationCaptured Sep 25, 2026
An account-wide activity log records record edits with old and new values, status and owner changes, exports and GDPR consent events, and can be filtered by user and date.
- Public statementCaptured Sep 25, 2026
The AI hub markets autonomous agents that screen applications and update the pipeline 'without manual intervention'. That conflicts with the help centre's position that a person makes every decision.
Post-Market Monitoring
6 items45
- Product featureCaptured Sep 25, 2026
An UptimeRobot status page at status.recruitly.io, linked as 'System Status' in the site footer, tracks four monitors with 90-day uptime (99.716% for the main application) but has no written incident updates.
- Product featureCaptured Sep 25, 2026
A second status page on Atlassian Statuspage (recruitly.statuspage.io) lists only one incident, 'This is an example incident' from 28 Dec 2023, so there is no public incident history.
- DocumentationCaptured Sep 25, 2026
A dated public changelog covers releases from March to September 2026, but no entry records a change to the models behind AI Match or AI Application Review scores.
- DocumentationCaptured Sep 25, 2026
The benchmarks page publishes production latency, context sensitivity and day-to-day drift for the models Recruitly runs. This is public monitoring of model performance, not of fairness or hiring outcomes.
- Public statementCaptured Sep 25, 2026
Recruitly logs every AI decision with its latency and serving host, and keeps a fallback path. It also says 'For recruiters none of this is visible, which is the intent', so customers are not told when a scoring model changes.
https://recruitly.io/resources/decision-model-in-production-the-day-it-was-released
- DocumentationCaptured Sep 25, 2026
Vulnerability reports go to the general support@recruitly.io address under a 'responsible disclosure programme' that has no published policy.
Customer Documentation
5 items58
- DocumentationCaptured Sep 25, 2026
A US guide (25 May 2026) explains NYC Local Law 144's audit, summary and ten-business-day notice duties and tells agencies to get a bias audit for any AEDT that touches NYC residents. It does not say whether Recruitly's own features count as AEDTs, and it still describes Colorado's SB 24-205 as taking effect on 30 June 2026.
- DocumentationCaptured Sep 25, 2026
Help-centre articles, updated July to September 2026, explain AI Matches, the AI review, screening policies, scorecards and the activity log step by step.
- DocumentationCaptured Sep 25, 2026
The GDPR automation page covers the consent lifecycle, archiving or anonymising records based on consent status, compiling subject access requests, and consent audit trails.
- DocumentationCaptured Sep 25, 2026
A published SLA commits to 99.9% monthly availability, with service credits of 10% to 100% and 48 hours' notice of scheduled maintenance.
- AbsenceCaptured Sep 25, 2026
No published DPA or sub-processor list, AI-specific candidate notice template, or product-specific statement of EU AI Act or LL 144 status was found. recruitly.io/dpa, /data-processing-agreement and /sub-processors return 404.
Editorial notes
Company overview
Recruitly is an agency ATS and CRM, founded in 2015 according to Tracxn and Latka. It started in Manchester. Its About page says it is 'born in the UK, run from Dubai' and that SOO Group moved to Dubai in 2025, with leadership now based there. The current terms say the service is operated by Recruitly, a division of The SOO Group, registered in Dubai (no. 2426166). An archived January 2024 copy of the terms named Recruitly UK Ltd as operator, and Companies House shows that company was dissolved on 25 November 2025. The site names Gowri as founder and CEO. AI Candidate Match scores candidates against every live job and shows reasons. It uses Qdrant, OpenAI embeddings and Claude. AI Application Review gives each applicant an Overall, Skills and Language score with a written explanation. An engineering post says TypeSafe's Jev model now scores skills, experience, location, visa and deal breakers during candidate review. Plans run from free to $120 per user per month.
Regulatory exposure
AI Match and AI Application Review score and rank candidates for agency hiring. That is high-risk AI under EU AI Act Annex III point 4, with Recruitly as provider and its agencies as deployers. The Annex III obligations now apply from 2 December 2027 after the Digital AI Omnibus deferral. They reach a Dubai-run vendor wherever its outputs are used in the EU. Agencies that use these scores to substantially assist hiring decisions for New York City roles need an annual independent bias audit and candidate notices under Local Law 144, and Recruitly publishes no audit. Illinois HB 3773 has applied since 1 January 2026. It requires notice when AI is used and bans discriminatory AI use and zip-code proxies. Colorado's SB 26-189 disclosure regime starts on 1 January 2027. California's FEHA ADS rules require four-year record retention, which agencies must configure themselves. UK GDPR and the ICO's automated-decision expectations cover its UK customer base. Recruitly's own regulatory guides misstate key facts. They place recruitment AI under Article 6(1) from August 2027 and describe Colorado's repealed SB 24-205 as taking effect on 30 June 2026.
Path to a higher score
The largest gain would be an independent bias audit of AI Application Review and AI Match scores. It should publish impact ratios by sex and race/ethnicity so agency customers can meet LL 144. Recruitly should also publish the AI position its own guides promise on its /ai page. That means a system card for each scoring feature covering intended purpose, inputs (including the culture-fit, language and visa signals), known limits, the evidence behind the '90% agreement' claim, and a self-assessment against Annex III. Other steps: publish the DPA and a sub-processor list naming model providers and routers, and reconcile the conflicting AWS, GCP and Cloudflare hosting statements. Add per-jurisdiction candidate-notice and alternative-process settings, and block protected-characteristic criteria in screening policies. Record model changes in the changelog and real incidents on a single status page, and correct the AI Act and Colorado dates in its guides. An ISO 27001 or ISO/IEC 42001 certificate held by Recruitly itself would add assurance.
Regulatory frame
What applies to ATS-integrated AI.
Ranking features are inside EU AI Act Annex III §4 (high-risk) and inside NYC Local Law 144 when they affect NYC-based candidates. Customers should treat the AI surface as a distinct deployer-FRIA scope from the rest of the ATS.
Compare
Recruitly against its nearest-scoring peers.
The closest scores in ATS-integrated AI, side by side.
- Side-by-sideRecruitly vs Cezanne Recruitment41F
- Side-by-sideRecruitly vs Harri41F
- Side-by-sideRecruitly vs Radancy41F
- Side-by-sideRecruitly vs Manatal41F
Ranking
Others rated in ATS-integrated AI
Ranked 61 of 77 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.
- #58Zoho Recruit42F
- #59Cezanne Recruitment41F
- #60Harri41F
- #61RecruitlyThis profile41F
- #62Radancy41F
- #63Manatal41F
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.