micro1
HireAIScore rates micro1 36 out of 100 (grade F, Substantial gaps) under rubric v1.0, last reviewed September 25, 2026. micro1 ranks 145 of 160 vendors rated in Screening AI, against a category average of 46. That score is drawn from 33 evidence items across 7 rubric criteria for micro1 — 25 with a cited source, 8 recording that nothing was located.
Score breakdown
micro1, 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 49
50
+10.0
Bias Audit Transparency
Weight 18% · category avg 41
20
+3.6
FRIA Support
Weight 15% · category avg 31
22
+3.3
Data Governance Disclosure
Weight 15% · category avg 53
45
+6.8
Human Oversight Design
Weight 12% · category avg 56
48
+5.8
Post-Market Monitoring
Weight 12% · category avg 40
25
+3.0
Customer Documentation
Weight 8% · category avg 58
45
+3.6
Total weighted score
36.0(rounded to 36 for display)
Category avg is the mean raw score on that criterion across the 160 Screening AI vendors in scope of this rubric, this one included.
Strongest criterion
Article 11 Technical Documentation
50raw / 100
Contributes 10.0 to the total. 50 against a 49 category average.
Weakest criterion
Bias Audit Transparency
20raw / 100
Contributes 3.6 to the total. 20 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
- 33
- Documentation
- 22
- Public statement
- 3
- Absence
- 8
- With a source URL
- 25 of 33
- Source hostnames
- 5
- Fewest items
- 3
- FRIA Support
These figures measure how thoroughly micro1 was reviewed, not how micro1 performed — an absence row, recording that nothing was located, is counted like any other item.
Article 11 Technical Documentation
5 items50
- DocumentationCaptured Sep 25, 2026
micro1's research page (April 2025) says Zara pairs GPT-4o with retrieval-augmented generation in four phases: an on-the-fly mock interview, a 20-40 minute branching AI interview, a structured skills report with a CEFR-style soft-skill rubric, and RAG-based answers to candidate emails. It says recruiters see the scores, transcript, video recording and feedback.
https://www.micro1.ai/research/zara-an-llm-based-candidate-interview-feedback-system
- DocumentationCaptured Sep 25, 2026
The matching arXiv paper by micro1 authors publishes the chain-of-thought prompts used to write candidate feedback and reports usage metrics, but it contains no fairness evaluation and no limitations section.
- DocumentationCaptured Sep 25, 2026
The field-experiment paper co-written by Stanford and micro1 lists what the AI Interview Report contains: four-level per-skill ratings, a 0-100 composite score, an automated proctoring score and a video/transcript link. It also gives the pass-label thresholds and an external check in which one developer hired through Upwork agreed with the AI on 96.1% of 180 skill ratings. It notes that the prompts and aggregation procedures are proprietary to micro1.
- DocumentationCaptured Sep 25, 2026
A micro1 study compares the speech-to-text, LLM and text-to-speech stacks behind Zara on more than 300,000 interview transcripts in eleven languages. It keeps OpenAI TTS fixed and swaps the speech engine (Google STT or Whisper 3) and the LLM (GPT-4.1, GPT-4o or Grok-2).
- AbsenceCaptured Sep 25, 2026
I found no model card, system card, instructions for use, explainability statement, AI principles page or ISO 42001 reference. I checked a full crawl of the micro1.ai sitemap, the Legal Center, the research index, the AI Recruiter API docs and the micro1 library.
Bias Audit Transparency
5 items20
- AbsenceCaptured Sep 25, 2026
I found no NYC LL 144 bias audit summary and no independent fairness audit of Zara. I checked micro1.ai (Legal Center, research index, full sitemap crawl), the AI Recruiter API docs and auditor portals: Warden AI returns 404, and the Holistic AI and Vanta URLs return only generic shells.
- DocumentationCaptured Sep 25, 2026
The Candidate Privacy Policy (updated July 21, 2026) tells NYC candidates they may request an alternative selection process or accommodation under Local Law 144 and says micro1 does not evaluate protected characteristics, but it links to no bias audit results.
- DocumentationCaptured Sep 25, 2026
The only published subgroup analysis is in the co-written field study. It infers gender from names and finds that men were 1.3-1.5 percentage points less likely than women to drop out of the AI interview. It reports no selection rates or impact ratios for Zara's scores or pass labels by sex or race/ethnicity.
- Public statementCaptured Sep 25, 2026
micro1's knowledge base for experts claims Zara interviews free of biased filtering but cites no audit or test results to support this.
- Public statementCaptured Sep 25, 2026
A September 2025 post by micro1's chief economist and CEO argues that AI hiring depends on rigorous validation and continuous bias auditing, yet micro1 publishes no such audit of Zara.
FRIA Support
3 items22
- AbsenceCaptured Sep 25, 2026
I found no EU AI Act Article 27 FRIA template, deployer-obligation guidance, high-risk classification statement or any EU AI Act reference. I checked micro1.ai, its Legal Center, the AI Recruiter API docs and the micro1 library.
- DocumentationCaptured Sep 25, 2026
The candidate privacy policy covers EU/UK data-subject rights, a right to opt out of automated decision-making and disclosures about transfers to the US, but says nothing about AI Act risk classification or impact assessments.
- AbsenceCaptured Sep 25, 2026
The employer-facing Zara pages that could have held deployer guidance (micro1.ai/zara, /zara/pricing, /zara/hire) now redirect to micro1's data-lab homepage, and the Legal Center lists no client terms or DPA.
Data Governance Disclosure
5 items45
- DocumentationCaptured Sep 25, 2026
The Candidate Privacy Policy lists the Zara interview data collected (audio/video recordings, screen-share recordings, transcripts, scores and proctoring analytics) plus optional government ID and photos. It also lists the uses, including generating assessment and proctoring scores.
- DocumentationCaptured Sep 25, 2026
The policy says de-identified interview audio, video, transcripts and results may be used to train micro1's own models, and that LLM vendors process data on micro1's behalf but never to train their models. Candidates can opt out of AI training.
- DocumentationCaptured Sep 25, 2026
Retention is described only in general terms (kept as long as the service needs it, with deletion on request) and no period is given for interview recordings.
- DocumentationCaptured Sep 25, 2026
The Candidate Terms (July 21, 2026) require candidates to consent to recording of their answers, audio, video and platform activity and to sharing of recordings and results, and they assign ownership of interview results to micro1.
- AbsenceCaptured Sep 25, 2026
I found no security page, trust center, subprocessor list, or SOC 2 / ISO 27001 / ISO 42001 attestation. micro1.ai/security and /trust return 404, trust.micro1.ai and security.micro1.ai do not resolve, and there is no micro1 portal on SafeBase, Vanta or Warden AI.
Human Oversight Design
7 items48
- DocumentationCaptured Sep 25, 2026
The automated decision-making section says AI screens applications, proctors interviews and evaluates candidacy, and that a human evaluator reviews the overall assessment before a final decision. Candidates can opt out of automated decision-making or appeal through privacy@micro1.ai.
- DocumentationCaptured Sep 25, 2026
Reports for recruiters include per-skill ratings with written AI feedback and timestamps into the interview, a timestamped transcript, the video recording, a proctoring score and itemized violations (tab_switch, eye_movement, ai_classifier).
https://ai-recruiter.micro1.ai/api-reference/interview-reports/get-all-completed-reports
- DocumentationCaptured Sep 25, 2026
Employers can configure each interview. They can turn proctoring off (it is on by default), set the interview language and allow candidates to change it, and add coding rounds or custom questions.
https://ai-recruiter.micro1.ai/api-reference/interview/create-a-new-interview
- DocumentationCaptured Sep 25, 2026
Candidate guidance says finishing the Zara interview produces a real-time result on whether the candidate meets micro1's global certification criteria, so the certification gate appears to be automated.
- DocumentationCaptured Sep 25, 2026
The experts FAQ says certification decisions are based on demonstrated knowledge, and that human recruiters review profiles and match experts to opportunities in addition to the automated systems.
- DocumentationCaptured Sep 25, 2026
The co-written field study describes an automatic pass label with fixed thresholds, including a proctoring score of at least 70%. It notes recruiters concentrate on candidates who passed, and that recruiters usually do not open the video/transcript link.
- Public statementCaptured Sep 25, 2026
micro1's March 28, 2025 launch press release pitches Zara to companies as running the whole recruitment process, from sourcing to onboarding, with minimal oversight.
Post-Market Monitoring
4 items25
- AbsenceCaptured Sep 25, 2026
I found no public status page. status.micro1.ai does not resolve, micro1.statuspage.io (including its components.json and incidents.json API endpoints) redirects to Atlassian's marketing site, and micro1.instatus.com and micro1.betteruptime.com redirect to the vendors' homepages.
- AbsenceCaptured Sep 25, 2026
I found no security.txt (micro1.ai/.well-known/security.txt returns 404), no vulnerability-disclosure policy or security contact, no incident channel and no model-update changelog for Zara. The only published addresses are the general support@ and privacy@ mailboxes.
- DocumentationCaptured Sep 25, 2026
The stack-comparison study shows micro1 tracks conversational and technical quality internally when it swaps Zara's speech and LLM components, but it is a single publication, not an ongoing public monitoring channel.
- DocumentationCaptured Sep 25, 2026
A September 2026 troubleshooting guide for candidates covers interviews that fail to load and escalation to support@micro1.ai, the only public channel for operational problems I found.
https://www.micro1.ai/experts-guide/interview-troubleshooting-guide
Customer Documentation
4 items45
- DocumentationCaptured Sep 25, 2026
The public AI Recruiter API reference (docs.micro1.ai redirects here) documents creating interviews, inviting candidates, retrieving reports and recordings, managing job applicants and webhooks, and publishes an OpenAPI spec.
https://ai-recruiter.micro1.ai/api-reference/getting-started/introduction
- DocumentationCaptured Sep 25, 2026
A guide for candidates explains that every candidate has a conversational interview with the AI recruiter Zara, describes the question types, and notes that candidates can request feedback.
- DocumentationCaptured Sep 25, 2026
The Candidate Privacy Policy gives candidates disclosures about AI and automated decision-making, appeal routes and an NYC Local Law 144 notice.
- AbsenceCaptured Sep 25, 2026
The employer-facing Zara product and pricing pages (micro1.ai/zara, /zara/pricing, /zara/hire) now redirect to micro1's data-lab homepage. The Legal Center lists only website, candidate and referral policies: no client terms, DPA, subprocessor list, or NYC/EU compliance guidance for deployers.
Editorial notes
Company overview
micro1, Inc. is a San Francisco-based company founded in 2022 by CEO Ali Ansari. It started as an AI recruiting startup and now presents itself mainly as a human-data lab that supplies expert data, RL environments and evaluations to frontier AI labs. Its AI interviewer, Zara, was announced in March 2025. Zara runs 20-40 minute adaptive spoken interviews with optional coding exercises and automated proctoring, then produces per-skill ratings (Not experienced to Senior), CEFR-style soft-skill ratings, a proctoring score and the full recording and transcript. Zara has two uses. First, every applicant must pass it to be certified into micro1's own expert network, the same model Mercor uses. Second, it was sold to outside employers with published pricing (a $399/month Growth plan) until early 2026. The employer marketing and pricing pages now redirect to the data-lab homepage, but the AI Recruiter API documentation and the client sign-up portal are still live.
Regulatory exposure
Zara evaluates and filters job candidates. Under the EU AI Act (Annex III(4)(a)) that makes it a high-risk system wherever it screens EU residents. micro1 is both provider and deployer when it certifies its global expert network, and provider to any employer using the API. For candidates who live in New York City it is an automated employment decision tool under Local Law 144. micro1's candidate privacy notice acknowledges the law by offering an alternative selection process, but micro1 publishes no bias audit summary. Recorded video interviews with eye-movement proctoring could bring in Illinois's AI Video Interview Act. Illinois's amended Human Rights Act (HB 3773) requires notice when AI is used in employment decisions. Colorado's AI Act (SB 24-205) treats employment as a consequential decision and places duties on both developers and deployers. micro1 also collects government ID, interview recordings and proctoring analytics, and trains its own models on de-identified interview data, which adds GDPR and US state privacy exposure.
Path to a higher score
Commission and publish an independent LL 144-style bias audit of Zara's certification pass decision and skill ratings. It should include impact ratios by sex, by race/ethnicity and intersectionally, cover the proctoring component, and be repeated every year. Publish a system card and instructions for use that cover the current model stack, scoring rubric, pass thresholds, supported languages and known limitations. Reconcile the claim that a human reviews each assessment with the real-time certification result, and document override, appeal and accommodation routes for all candidates, not only NYC residents. State retention periods for recordings and ID data. Stand up a trust center with security attestations, a subprocessor list and a DPA. Add EU AI Act deployer guidance and a FRIA template. Launch a public status page and a model-update changelog. Finally, either maintain or formally retire the employer-facing Zara documentation so customers know whether the product is still supported.
Regulatory frame
What applies to screening AI.
Treated as high-risk under Annex III §4 when the screening output is decisive. EU AI Act Article 13 transparency obligations apply; deployers must give candidates a way to contest.
Compare
micro1 against its nearest-scoring peers.
The closest scores in Screening AI, side by side.
- Side-by-sidemicro1 vs Fetcher37F
- Side-by-sidemicro1 vs TestDome37F
- Side-by-sidemicro1 vs myInterview37F
- Side-by-sidemicro1 vs Dice36F
Ranking
Others rated in Screening AI
Ranked 145 of 160 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.
- #142Keka37F
- #143Toggl Hire37F
- #144Dice36F
- #145micro1This profile36F
- #146Aon Assessment35F
- #147HireLogic34F
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.