pre-employment assessment toolsrankingsvendor evaluationbias audit

The best pre-employment assessment tools in 2026, ranked on evidence

58 pre-employment assessment tools ranked by HireAIScore on public evidence: bias audits, technical documentation, FRIA support, human oversight and monitoring.

By HireAIScore editorial · 14 min read

On this page 16 sections
  1. Why assessment is the most regulated surface
  2. How we ranked them
  3. The top ten at a glance
  4. 1. Maki People - 67/100
  5. 2. HiPeople - 61/100
  6. 3. Alva Labs - 61/100
  7. 4. Bryq - 57/100
  8. 5. Criteria - 56/100
  9. 6. Sapia.ai - 56/100
  10. 7. Harver - 56/100
  11. 8. iMocha - 56/100
  12. 9. SHL - 56/100
  13. 10. Plum - 54/100
  14. The rest of the category
  15. How to choose a pre-employment assessment tool
  16. Frequently asked questions

Key takeaways

  • Maki People (67), HiPeople (61) and Alva Labs (61) are the only pre-employment assessment tools in the directory to reach a D. The other 55 score F, and none reaches a C.
  • Audit coverage is the category's biggest gap: vendors often publish a bias audit for one product while selling several. HiPeople's monthly audits cover its application screening, not its test library.
  • Maki People is the only assessment vendor whose FRIA support scores above 42, and even its FRIA material is provided on request.
  • Several published audits are more than a year old or predate the scoring now in use, including Bryq's (2024 data), Harver's and Plum's (2025) and Sapia.ai's (2022).
  • Before buying, ask which products the audit covers, what data it used, and whether any setting rejects candidates automatically below a cut-off.

Pre-employment assessment tools - cognitive and aptitude tests, personality questionnaires, situational judgement tests, skills and coding tests, and increasingly AI-scored interviews - all produce the same thing: a score that decides who moves forward. That is why assessment is the most regulated surface in AI hiring, and why the evidence behind a score matters as much as the science behind a test.

HireAIScore scores every vendor on the same rubric, from public evidence only. Below are all 58 tools in the directory whose primary category is assessment, ranked by score. The score measures public evidence of compliance readiness: bias audits, documentation, oversight and monitoring. It does not measure whether a test is valid for your role or how candidates experience it.

Why assessment is the most regulated surface

An assessment score is used to select people, so almost every regime reaches it:

  • NYC Local Law 144. An assessment score used as a substantial factor in a hiring or promotion decision needs an independent bias audit from within the past year, a published summary and candidate notice. More on LL 144, including what counts as an AEDT.
  • EEOC and OFCCP frameworks for selection procedures also apply in the US.
  • Illinois HB 3773 has applied since 1 January 2026 and requires notice when AI is used in employment decisions. More on HB 3773.
  • Colorado's SB 26-189 covers automated decisions in employment from 1 January 2027. More on Colorado.
  • EU AI Act. Assessment used in recruitment is high-risk under Annex III §4. After the Digital Omnibus on AI, those obligations apply from 2 December 2027. Certain emotion recognition has been prohibited since February 2025. More on the EU AI Act.

These overlap most on bias audits. Many suites mix traditionally scored psychometric tests with AI-scored interviews, coding graders and proctoring, and an audit of one product says nothing about the others. That gap runs through most of the write-ups below.

How we ranked them

Each tool gets a 0-100 score in seven categories, weighted into one overall score:

CategoryWeightWhat it asks
Article 11 technical documentation20%Is there Annex IV-aligned documentation of how the system works?
Bias audit transparency18%Is there a published, recent bias audit, and what did it test?
FRIA support15%Could an employer complete a fundamental-rights impact assessment from public material?
Data governance disclosure15%Training data, subgroup performance, retention, data rights
Human oversight design12%Overrides, explanations, audit logs, the ability to stop the system
Post-market monitoring12%Incident channels, model changelogs, drift monitoring
Customer documentation8%Instructions for use, limitations, intended-use boundaries

The grade bands are the same for every vendor: 90 and above is an A, 80 a B, 70 a C, 60 a D, and anything lower an F. No assessment tool reaches a C, and three reach a D. Where tools share a rounded score, they are listed in order of their unrounded score.

The top ten at a glance

RankToolScoreBias auditFRIA supportLast reviewed
1Maki People67 (D)80622026-11-01
2HiPeople61 (D)84352026-08-02
3Alva Labs61 (D)58382026-08-02
4Bryq57 (F)70402026-07-17
5Criteria56 (F)52402026-11-14
6Sapia.ai56 (F)46382026-11-14
7Harver56 (F)68352026-11-14
8iMocha56 (F)44382026-07-17
9SHL56 (F)58422026-11-14
10Plum54 (F)72332026-11-14

The two middle columns are category scores out of 100. Five tools share a rounded 56; their unrounded scores separate them.

1. Maki People - 67/100

Maki People is a Paris-based platform for high-volume assessment and screening: skills, behavioural and situational tests, voice screening in more than 45 languages, and conversational agents for interview support and scheduling.

Strengths. The highest score in the category, and the only assessment vendor with FRIA support above 42: it gives customers a FRIA information deck, an EU AI Act deck and a DPIA summary built for deployers. Two publicly viewable Holistic AI LL 144 audits cover its Shiro assessment engine and its Mochi voice screening, each with full impact-ratio tables, and it commits to annual audits. Its recruiter manuals document a score override that requires a written reason, keeps the original score visible and records who changed it, plus a control to halt assessments. Every candidate scorecard includes an AI-use disclosure, and Maki states that no score can trigger automatic rejection.

Watch out for. The FRIA deck, Instructions for Use, DPIA summary and SOC 2 report are all provided on request, not published. There is no ISO 42001 certificate, no Annex IV-style technical pack or model card, and the audits rely on synthetic test data, with no live applicant-flow results yet.

Best for: high-volume employers, especially in the EU, who need deployer material for a FRIA and documented human override.

2. HiPeople - 61/100

HiPeople is a Berlin-based platform with a library of 400+ skills and personality tests, automated AI reference checks, AI application screening and AI screening calls.

Strengths. The highest bias-audit score in the category. Warden AI audits its AI Application Screening monthly, with a public dashboard (latest 9 July 2026 at review, 19,092 samples) and a published LL 144 report giving selection rates and impact ratios by sex, race/ethnicity and eight intersectional groups, all between 0.94 and 1.00.

Watch out for. That assurance covers application screening only. The assessment library, reference checks and AI screening calls - the products the category rests on - carry no published fairness evidence. There is no ISO 42001, no model or system card, no published validity evidence behind the tests and nothing on FRIA support for deployers.

Best for: teams using HiPeople's AI screening, who should ask for fairness evidence on the specific tests they plan to use. HiPeople's unrounded score is slightly above Alva Labs'.

3. Alva Labs - 61/100

Alva Labs, based in Stockholm, sells an adaptive logic test built on Item Response Theory, a Five-Factor personality test and coding tests, with candidate management around them.

Strengths. The best technical documentation in the category. Alva Labs is certified to ISO/IEC 42001 by DNV, covering its psychometric tests and platform AI, and holds ISO/IEC 27001 and ISO 10667-2. It publishes a feature-by-feature account of how AI is used, a downloadable technical manual for its logic test, and states that every test item goes through differential item functioning analysis before deployment. Candidate data stays in the EU and is not used to train third-party models.

Watch out for. No NYC LL 144 audit summary and no independent bias audit from a named auditor. Its product pages claim minimal group differences across gender, ethnicity and age without published impact ratios or sample sizes. There is no FRIA template and no NYC, Illinois or Colorado guidance.

Best for: EU employers who value certified AI governance. Employers hiring for NYC roles would need to commission their own audit.

4. Bryq - 57/100

Bryq runs a roughly 14-minute chatbot-driven assessment of cognitive ability, personality and job skills, and scores candidates on fit tiers against an Ideal Candidate Profile.

Strengths. A publicly downloadable LL 144 audit by code4thought (March 2025) on 9,846 real candidates from 2024, with gender, ethnicity and intersectional tables and a lowest impact ratio of 0.821. It has used Holistic AI audits since 2023, holds SOC 2 Type II and ISO 27001:2022, and publishes its DPA and subprocessor list.

Watch out for. The public audit covers 2024 data, and the Holistic AI reports are available only on request. There is no ISO 42001, no model card or Annex IV-style file, no FRIA template, and its human oversight is described as decision support without documented override or audit-log controls.

Best for: mid-market employers who want to read an audit on real candidates before they buy.

5. Criteria - 56/100

Criteria sells aptitude tests including the CCAT, personality, emotional-intelligence (Emotify) and skills tests, plus video interviews with an AI Scoring add-on, an AI Interview Agent rolling out from November 2026, and AI Proctoring.

Strengths. A public Legal Defensibility page has summarised an annual LL 144-style audit of its assessments since 2023; the June 2026 summary covers more than 300,000 candidates and reports no practically significant disparate impact by gender, race or their combination. Its help centre documents AI Scoring, Hybrid Scoring and proctoring controls.

Watch out for. The summary names no auditor and gives no impact-ratio tables, and AI Scoring, the AI Interview Agent and AI Proctoring have no published audit. On 6 October 2026 the ACLU asked California's Civil Rights Department to investigate Criteria, focusing on Emotify; that is a complaint, not a finding.

Best for: employers using its established aptitude tests who will request the full audit and keep AI interview scoring under review.

6. Sapia.ai - 56/100

Sapia.ai scores a five-to-seven-question, text-only chat interview - no video or audio - with an LLM-based engine, SAIGE.

Strengths. ISO/IEC 42001 certified (first issued September 2025 and kept after a September 2026 surveillance audit). SAIGE shows a plain-language explanation and the excerpt of the candidate's answer behind each score. Its help centre and DPA are public.

Watch out for. The only independent audit it publishes is a December 2022 summary from BLDS, which predates its LLM scoring and is an executive summary without sample sizes or impact-ratio tables. Its product page still says the interview is "independently audited for bias, published openly". Its AI Impact Assessment, DPIA and Bias Monitoring Policy are request-only.

Best for: employers who want an explainable, text-only interview and will ask for a current audit of SAIGE first.

7. Harver - 56/100

Harver covers cognitive, behavioural and situational-judgement assessments, video interviewing, the pymetrics game-based assessments and Checkster reference checks, and acquired Symphony Talent in September 2026.

Strengths. Separate public BABL AI LL 144 audits for the Harver Platform and pymetrics, a large public knowledge base, and since mid-2026 a dedicated address for reporting AI issues. It gave customers advance notice of a norm update that would shift pass rates, and publishes dated release notes.

Watch out for. Both published audits date from 2025, and the Harver Platform audit used two months of data on about 280 applicants with known gender. Harver says it passed a 2026 audit but has not published it. A May 2026 Stanford, Chapman and Northeastern study of 4.2 million pymetrics applications reported position-level racial disparities that pooled audits do not show; Harver disputes the study. The main platform has no public status page. Pymetrics is scored separately, at 50.

Best for: volume hirers who want assessments, video and reference checks from one vendor, and will ask for the 2026 audits.

8. iMocha - 56/100

iMocha is a skills assessment and skills-intelligence platform with coding simulators and Tara, an AI video interviewer, used for hiring and internal mobility.

Strengths. A public 15-page technical overview of its AI inference engine covers data sources, scoring, explainability and model governance, with a per-data-element retention table. It documents HR override with a recorded rationale and per-inference explanations, and states that its AI interview analyses spoken and written responses only, with no biometric or facial analysis.

Watch out for. No published independent bias audit: its bias-audit summaries are offered only on written request. There is no FRIA or deployer guidance, no ISO 42001 and no public status page or model changelog.

Best for: skills-based hiring where the employer will run its own adverse-impact testing.

9. SHL - 56/100

SHL sells traditionally scored cognitive, personality and behavioural tests alongside AI-scored video, coding, spoken-language, writing and chat-simulation assessments. It says it delivers about 45 million assessments a year.

Strengths. SHL lists 12 AI-scored assessments as in scope for LL 144 and says its standard audits are updated annually and free to customers. It publishes fairness reports with impact-ratio tables, a product-by-product AI overview, a public help centre and an EU AI Act hub with a September 2026 deployer checklist. Its FRIA support (42) is joint second in the category.

Watch out for. Its LL 144 audits go to customers rather than being published; public summaries exist for Automata and two Skills Insight assessments, none naming the auditor. Its current Smart Interview On Demand guide describes optional AI scoring of facial expressions and voice tonality, which conflicts with its AI Policy and would need checking against the EU AI Act's prohibition on certain emotion recognition.

Best for: large employers who want an established psychometric provider and will obtain the audit for each product they use.

10. Plum - 54/100

Plum scores candidates 30-99 on role fit from a Five-Factor personality survey. Phenom announced its acquisition in April 2026 and now feeds Plum's behavioural match into its Sourcing Agent.

Strengths. A multi-year record of published LL 144 audits: FairNow in October 2025 on about 557,000 applications, FairNow in 2024 and Holistic AI in 2023. The 2025 addendum also reports age and disability ratios, which LL 144 does not require. A public help-centre Trust Center explains how its models are built.

Watch out for. The latest audit is now more than a year old, and its own supplemental full-sample test puts Black and Black female candidates just under the four-fifths threshold. No audit covers Phenom's use of the scores in agentic sourcing, and the technical manual and validity evidence are gated.

Best for: employers who want personality-based matching with a long audit record, and will check how the move to Phenom changes it. See the AI sourcing tools ranking for Phenom itself.

The rest of the category

RankToolScoreLast reviewed
11HackerRank54 (F)2026-11-14
12Codility54 (F)2026-07-17
13HiringBranch52 (F)2026-07-17
14Glider AI52 (F)2026-09-25
15TestGorilla51 (F)2026-11-14
16Pymetrics (Harver Games)50 (F)2026-11-14
17Xobin50 (F)2026-09-25
18Test Partnership50 (F)2026-09-25
19Korn Ferry49 (F)2026-11-14
20Versant by Pearson49 (F)2026-09-25
21Infor Talent Science49 (F)2026-09-25
22Vervoe49 (F)2026-11-14
23X0PA AI48 (F)2026-07-17
24Sova Assessment48 (F)2026-08-02
25The Predictive Index48 (F)2026-07-17
26Saville Assessment48 (F)2026-09-25
27Thomas International48 (F)2026-07-17
28Hogan Assessments47 (F)2026-07-17
29CodeSignal47 (F)2026-11-14
30Wonderlic47 (F)2026-11-14
31HackerEarth46 (F)2026-09-25
32TalentLens46 (F)2026-09-25
33Retorio46 (F)2026-11-14
34Arctic Shores46 (F)2026-08-02
35Testlify45 (F)2026-09-25
36Mercer | Mettl45 (F)2026-11-14
37Clevry45 (F)2026-09-25
38Cangrade45 (F)2026-08-02
39Equalture44 (F)2026-08-02
40Emmersion44 (F)2026-09-25
41Prevue HR44 (F)2026-09-25
42HighMatch43 (F)2026-09-25
43Coderbyte42 (F)2026-09-25
44Talogy42 (F)2026-08-02
45CoderPad42 (F)2026-09-25
46eSkill41 (F)2026-09-25
47Feenyx40 (F)2026-09-25
48Aon Assessment40 (F)2026-11-14
49Cappfinity40 (F)2026-09-25
50Adaface39 (F)2026-09-25
51Canditech39 (F)2026-07-17
52DevSkiller39 (F)2026-07-17
53Alooba38 (F)2026-09-25
54TestDome37 (F)2026-09-25
55Toggl Hire37 (F)2026-07-17
56Turing36 (F)2026-07-17
57Pulsifi33 (F)2026-09-25
58Cognisess30 (F)2026-09-25

The assessment category page also lists tools whose primary category is something else but which sell assessments, such as HireVue at 62 (D), ranked among HireVue alternatives, and Metaview at 60 (D). For every employer-side category in one place, see the best AI recruiting tools.

How to choose a pre-employment assessment tool

The score tells you how much a vendor has shown. These questions tell you whether it fits how you will use it:

  1. Which products does the audit cover? Ask for a list of every scored product in the suite - tests, AI interview scoring, coding graders, proctoring flags - and which ones the bias audit tested. An audit of one product says nothing about the rest.
  2. What is behind the audit? Ask for the auditor's name, the data window, sample sizes, intersectional results and whether it used real candidates or a test set. A two-month window or a few hundred applicants is thin. See what makes an LL 144 audit defensible.
  3. Is the test valid for this role? Ask for the technical manual and validity evidence for jobs like yours, not just a general claim of science.
  4. Can a recruiter override a score? Look for an override that records a reason and who made it, and check whether any setting rejects candidates automatically below a cut-off.
  5. What exactly is measured? If video or voice is analysed, ask whether facial expressions or tone of voice feed the score. In the EU, certain emotion recognition is prohibited.
  6. What do EU deployers get? Ask for instructions for use and FRIA inputs, and check them against the nine sections of Annex IV. The EU AI Act quick check tells you which obligations reach your use.

Every score above links to a full scorecard with the evidence behind each category. If you also source candidates with AI, see the AI sourcing tools ranking. Scores change when vendors publish new evidence, which they can submit for review.

Frequently asked questions

What are the best pre-employment assessment tools in 2026?

On HireAIScore's rubric, Maki People scores highest at 67/100 (D), followed by HiPeople and Alva Labs at 61 (D) and Bryq at 57 (F). No pre-employment assessment tool in the directory reaches a C, so the ranking is least concerning first, not a list of endorsements.

Are pre-employment assessments covered by AI hiring laws?

Assessment is the most regulated surface in AI hiring. NYC Local Law 144, EEOC and OFCCP frameworks, Illinois HB 3773 and EU AI Act Annex III §4 all apply, with overlap on bias-audit obligations. Colorado's SB 26-189 adds duties from 1 January 2027, and the EU AI Act's high-risk obligations apply from 2 December 2027.

Do pre-employment assessment vendors publish bias audits?

Many do, but often for only part of what they sell. HiPeople's monthly Warden AI audits cover its application screening, not its tests; Maki People publishes Holistic AI audits of two products; Bryq, Harver and Plum publish audits on 2024 or 2025 data; Alva Labs and iMocha publish no independent bias audit.

What should I ask a pre-employment assessment vendor before buying?

Which products are scored by AI and which the bias audit covers; the auditor, data window, sample sizes and whether the audit used real candidates; the technical manual and validity evidence for your role; whether a recruiter can override a score with a logged reason; whether any setting rejects candidates automatically; and what it gives EU deployers for a FRIA.

How were these assessment tools ranked?

Each tool is scored from public evidence on HireAIScore's rubric v1.0: Article 11 technical documentation (20%), bias audit transparency (18%), FRIA support (15%), data governance disclosure (15%), human oversight design (12%), post-market monitoring (12%) and customer documentation (8%). The ranking is the overall score.