Retorio
HireAIScore rates Retorio 44 out of 100 (grade F, Substantial gaps) under rubric v1.0, last reviewed June 7, 2026. Retorio ranks 51 of 82 vendors rated in Assessment AI, against a category average of 46. That score is drawn from 22 evidence items across 7 rubric criteria for Retorio — 16 with a cited source, 6 recording that nothing was located.
Score breakdown
Retorio, 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 50
50
+10.0
Bias Audit Transparency
Weight 18% · category avg 41
40
+7.2
FRIA Support
Weight 15% · category avg 31
30
+4.5
Data Governance Disclosure
Weight 15% · category avg 53
52
+7.8
Human Oversight Design
Weight 12% · category avg 55
52
+6.2
Post-Market Monitoring
Weight 12% · category avg 38
33
+4.0
Customer Documentation
Weight 8% · category avg 58
50
+4.0
Total weighted score
43.7(rounded to 44 for display)
Category avg is the mean raw score on that criterion across the 82 Assessment AI vendors in scope of this rubric, this one included.
Strongest criterion
Human Oversight Design
52raw / 100
Contributes 6.2 to the total. 52 against a 55 category average.
Weakest criterion
FRIA Support
30raw / 100
Contributes 4.5 to the total. 30 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 recordEvidence ledger
- Items
- 22
- Documentation
- 11
- Audit report
- 2
- Public statement
- 2
- Product feature
- 1
- Absence
- 6
- With a source URL
- 16 of 22
- Source hostnames
- 4
- Fewest items
- 2
- FRIA Support
These figures measure how thoroughly Retorio was reviewed, not how Retorio performed — an absence row, recording that nothing was located, is counted like any other item.
Article 11 Technical Documentation
3 items50
- DocumentationCaptured Jun 7, 2026
Science page describing validation against UCLA FairFace and a named XAI 'reasoning model' explainability layer, but linking no model cards or papers.
- DocumentationCaptured Jun 7, 2026
Dedicated EU AI Act page laying out three principles and referencing EU trustworthy-AI guidelines, with no Article 11 technical documentation.
- AbsenceCaptured Jun 7, 2026
No public material located as of 2026-06-07.
Where we looked: No ISO 42001, published model card, or downloadable technical/conformity pack located across retorio.com.
Bias Audit Transparency
4 items40
- Public statementCaptured Jun 7, 2026
States its AI was 'debiased and tested on global datasets, incl. the UCLA FairFace dataset'- a self-reported internal test, not an independent audit, with no downloadable results.
- Audit reportCaptured Jun 7, 2026
MIT Technology Review reports the 2021 Bavarian Broadcasting (BR) investigation found Retorio's scores changed with glasses and video background.
https://www.technologyreview.com/2021/07/07/1027916/we-tested-ai-interview-tools/
- Audit reportCaptured Jun 7, 2026
Independent coverage detailing specific OCEAN-score shifts from glasses, a headscarf, a bookshelf background, and video brightness.
- AbsenceCaptured Jun 7, 2026
No public material located as of 2026-06-07.
Where we looked: No independent third-party bias audit (BABL/DCI/ORCAA/Warden/Holistic/Credo) located, No NYC LL 144 AEDT audit located.
FRIA Support
2 items30
- DocumentationCaptured Jun 7, 2026
AI Act page provides a general compliance self-assessment but no Article 27 FRIA template or deployer-obligation walkthrough.
- AbsenceCaptured Jun 7, 2026
No public material located as of 2026-06-07.
Where we looked: No FRIA template or deployer FRIA guidance located; the AI Act blog argues the product is out of scope rather than supporting deployers.
Data Governance Disclosure
4 items52
- DocumentationCaptured Jun 7, 2026
Ethical-data page states the system excludes skin color, gender, and age, performs no biometric identification or emotion recognition, hosts data on ISO-27001 EU servers, and runs quarterly penetration tests.
- DocumentationCaptured Jun 7, 2026
States 'Retorio's base models do not learn from client data' and train on proprietary plus public scientific datasets (FairFace).
- DocumentationCaptured Jun 7, 2026
Privacy policy names the controller, an external DPO (DataCo GmbH), a subprocessor table, and short log-retention windows.
- AbsenceCaptured Jun 7, 2026
No public material located as of 2026-06-07.
Where we looked: No SOC 2, ISO 42001, or consolidated security/trust page located; privacy policy silent on training-use of candidate video.
Human Oversight Design
3 items52
- DocumentationCaptured Jun 7, 2026
States 'there is no automated decision making, the system is not self-learning or automatically updating,' and that users are informed an AI system is in use.
- DocumentationCaptured Jun 7, 2026
Describes in-product explainability - timestamp-linked graphs, keyword highlighting, and textual XAI explanations.
- Public statementCaptured Jun 7, 2026
Co-founder states the product is 'an assisting tool... not an automatic filter,' framing it as human-in-the-loop decision support.
https://www.technologyreview.com/2021/07/07/1027916/we-tested-ai-interview-tools/
Post-Market Monitoring
2 items33
- DocumentationCaptured Jun 7, 2026
Mentions quarterly penetration testing and disaster-recovery plans, but no continuous-monitoring dashboard or model-update changelog.
- AbsenceCaptured Jun 7, 2026
No public material located as of 2026-06-07.
Where we looked: No public status page, security.txt, vulnerability-disclosure policy, trust center, or model-update changelog located.
Customer Documentation
4 items50
- DocumentationCaptured Jun 7, 2026
Public FAQ covering EU AI Act stance, ISO 27001, data deletion, bias mitigation, and non-surveillance design.
- DocumentationCaptured Jun 7, 2026
Dedicated AI Act resource page consolidating Retorio's compliance positioning for buyers.
- Product featureCaptured Jun 7, 2026
Third-party (Greenhouse) integration documentation confirming Retorio's deployment as a recruiting/assessment tool within an ATS.
https://support.greenhouse.io/hc/en-us/articles/360028914231-Retorio-integration
- AbsenceCaptured Jun 7, 2026
No public material located as of 2026-06-07.
Where we looked: No downloadable DPA, NYC LL 144 candidate-notice guidance, or EU AI Act deployer guidance located; the /personality-assessment product page returns HTTP 404.
Editorial notes
Company overview
Retorio is a Munich-based behavioural-AI company spun out of the Technical University of Munich in 2018. Its core technology analyses short videos across three channels - language, voice, and facial expression/body language - to produce Big Five (OCEAN) personality and behavioural profiles. The product was first sold for AI video recruiting and candidate pre-screening and has since been repositioned as an enterprise "Behavioral Intelligence" coaching platform; the standalone personality-assessment product page now returns a 404, part of a deliberate move to argue the tool is not a high-risk Annex III system.
Regulatory exposure
As an EU vendor performing video-based personality/behaviour inference, Retorio sits in one of the EU AI Act's most contested zones. The company publicly claims to be "100% AI Act compliant" and not high-risk, resting on a self-serving reading that it performs "behavioural analysis" rather than prohibited emotion recognition, and that a "coaching" tool falls outside Annex III. Critically, in 2021 a Bayerischer Rundfunk investigation (covered by MIT Technology Review) showed Retorio's OCEAN scores changing materially when an actress added glasses, a headscarf, or a bookshelf background, or when video brightness changed - a documented fairness concern the company has never resolved with an independent audit.
Path to a higher score
Retorio's compliance posture is built on self-assertion rather than verifiable third-party evidence. To raise its scores it would need to commission and publish an independent bias/fairness audit that addresses the 2021 BR findings directly; pursue ISO 42001 and publish Article 11-style technical documentation and a model card; provide explicit EU AI Act deployer guidance and a FRIA template; and stand up a public trust center with a security-disclosure channel and model-update changelog.
Regulatory frame
What applies to assessment AI.
Assessment is the most-regulated surface in AI hiring. NYC LL 144, EEOC and OFCCP frameworks, Illinois HB 3773, and EU AI Act Annex III §4 all apply, with overlap on bias-audit obligations.
Compare
Retorio against its nearest-scoring peers.
The closest scores in Assessment AI, side by side.
- Side-by-sideRetorio vs Testlify45F
- Side-by-sideRetorio vs Equalture44F
- Side-by-sideRetorio vs Emmersion44F
- Side-by-sideRetorio vs Prevue HR44F
Ranking
Others rated in Assessment AI
Ranked 51 of 82 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.
- #48Equalture44F
- #49Emmersion44F
- #50Prevue HR44F
- #51RetorioThis profile44F
- #52HighMatch43F
- #53Wonderlic43F
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.