Headquarters: Melbourne, Australia, with a London office
Assessed: Structured chat interview scored by machine-learning and natural-language models.
Rubric: version 1. Search protocol applied and all sources accessed: 2026-09-22.
Documented 2, partial 3, absent 3, of eight criteria. This records what is published about the validity and fairness of the scores. It is not an assessment of the product, the model or the underlying science.
| # | Criterion | Level |
|---|---|---|
| C1 | Technical manual or validity report publicly available without a sales conversation | P |
| C2 | Validity coefficients reported with sample size and uncertainty | A |
| C3 | Adverse impact reported by named subgroup | P |
| C4 | Scored construct named and defined, not described only by outcome | D |
| C5 | Evidence of comparability across languages, versions or delivery modes | A |
| C6 | Human oversight documented, with a candidate route to contest a score | P |
| C7 | Independent review named and the reviewer identified | D |
| C8 | Reliability reported, and reported by group | A |
C1. Technical manual or validity report publicly available without a sales conversation
Partial.
A whitepaper, "Towards establishing fairness in AI-based candidate screening", is offered through a landing page rather than as a direct download.
Source: https://sapia.ai/resources/whitepaper/towards-establishing-fairness-in-ai-based-candidate-screening/ (accessed 2026-09-22).
The document requires contact details and was not read.
C2. Validity coefficients reported with sample size and uncertainty
Absent.
No public source was found for this criterion by the search protocol.
Pages examined include https://sapia.ai/candidate-explainer/ (accessed 2026-09-22).
C3. Adverse impact reported by named subgroup
Partial.
An independent disparate impact report is published, conducted with "the Standardized Mean Difference (SMD) and the Adverse Impact Ratio (AIR)", in which "BLDS performed a total of 23 tests for protected groups on the North American models".
Source: https://sapia.ai/candidate-explainer/ (accessed 2026-09-22).
The method, the metrics and the number of tests are stated, and the groups are described as protected groups. Per-group figures were not found on the vendor pages examined.
C4. Scored construct named and defined, not described only by outcome
Documented.
"HEXACO is a widely used and accepted six trait model of personality that conceptualises personality in terms of these six dimensions", with Honesty/Humility, Emotionality, Extraversion, Agreeableness, Conscientiousness and Openness to Experience each listed with sub-traits.
Source: https://sapia.ai/candidate-explainer/ (accessed 2026-09-22).
A published personality model is named, its dimensions are enumerated with sub-traits, and the explanation sits on a candidate-facing page rather than a buyer-facing one.
C5. Evidence of comparability across languages, versions or delivery modes
Absent.
No public source was found for this criterion by the search protocol.
Pages examined include https://sapia.ai/candidate-explainer/ (accessed 2026-09-22).
C6. Human oversight documented, with a candidate route to contest a score
Partial.
The candidate page explains that results are "based on your responses to the Chat Interview questions" and offers a help route for technical issues.
Source: https://sapia.ai/candidate-explainer/ (accessed 2026-09-22).
The candidate is told what is scored and on what basis. No route to query or contest the score itself was found.
C7. Independent review named and the reviewer identified
Documented.
An independent bias audit is published and the auditor is named: the disparate impact report by BLDS, LLC.
Source: https://sapia.ai/candidate-explainer/ (accessed 2026-09-22).
A named external auditor with a stated method. The vendor also states an ISO/IEC 42001 certification, which is a management-system standard rather than a review of measurement quality.
C8. Reliability reported, and reported by group
Absent.
No public source was found for this criterion by the search protocol.
Pages examined include https://sapia.ai/candidate-explainer/ (accessed 2026-09-22).
Absent records the outcome of the published search protocol on 2026-09-22. Where a source is supplied that the search did not reach, the cell is re-scored under the corrections policy, with the date of the change recorded and the previous level retained.
Back to the index, the rubric and the methods note