Micro1

Micro1 offers Zara, an AI-powered interviewer for structured, asynchronous candidate interviews. The product uses large language models to analyse responses, score them against predefined criteria and produce structured results for recruiters. Micro1 explicitly describes the process as human in the loop: people are intended to remain involved at every decision point. Zara is therefore particularly relevant to organisations that want consistency across repeatable first-round interviews without delegating the hiring recommendation to an opaque system. Source: Micro1, Who is Zara?, checked 20 August 2026.

Micro1 is no longer positioned only as a recruiting company. Its corporate website presents it as a data lab for expert human data, training environments and AI evaluations, while Zara is described separately as an AI recruiter. That distinction matters in procurement: the proposed Zara deployment for a specific recruiting workflow, the talent marketplace and other Micro1 services should be clearly separated in the commercial and data-processing scope. This directory profile assesses only Zara’s publicly documented recruiting functionality. Source: Micro1 corporate site and product documentation, checked 20 August 2026.

Micro1 and Zara: facts at a glance

What Zara does technically in a recruiting workflow

Zara is more than a static question-set tool. A team can create an interview against an open job or as a standalone assessment, select an interview type, define skills and custom questions, and configure language, exercises and proctoring. The standard format supports up to five skills and up to ten custom questions. For résumé-based interviews, Zara can analyse an uploaded résumé or one imported through an ATS connection and tailor the conversation to the candidate’s skills and experience. This represents a different automation level from a fixed one-way video questionnaire: the system is intended to respond within an employer-defined assessment framework, while staying bounded by the role and criteria the employer supplies. Source: Micro1, Types of Interviews and Create Interview, checked 20 August 2026.

The documented data sources are broad. Depending on the workflow, Zara processes job information and skills, résumé data, spoken interview answers, coding or custom-exercise results, language signals and, if proctoring is enabled, behavioural and environmental data. Its AI Match Score ranges from 0% to 100% and can combine résumé analysis, structured interviews, live problem-solving exercises and soft-skills evaluation. Micro1 says employers can describe in advanced settings how the score should be weighted or interpreted. That can be useful for roles with clear, defensible criteria; it does not remove the need to test whether those criteria are job-relevant, non-discriminatory and appropriate for every candidate. Source: Micro1, AI Match Score, checked 20 August 2026.

Once an interview is complete, Zara generates a report with a score and detailed explanations across skills, questions, exercises and soft skills. A report can be downloaded as a PDF or shared through a public URL, so the employer must ensure that only authorised people receive access. Video and transcript are also available. Micro1 says a low-resolution video is available immediately and a high-resolution version may take up to 30 minutes to process. It also notes that speech clarity, audio quality and background noise can affect transcript quality. This limitation matters whenever a score or a consequential recommendation relies on spoken answers. Source: Micro1, Interview Report, checked 20 August 2026.

Automation can range from manual to largely automatic. Candidates may be invited with a link, individually from the dashboard, through a CSV upload or from an ATS connection. For a job, automatic invitations can be enabled for new and uploaded applicants. Micro1 also documents application intake, pipeline stages, résumé-based ranking and publishing to job boards. Zara can therefore serve as a lightweight ATS for early recruitment stages. Whether it can replace an established ATS, however, depends not on the existence of an API alone but on field mappings, status write-backs, permissions, retention controls and the exact integration. The public documentation does not describe those details for individual target systems. Source: Micro1, All-in-one ATS platform, Invite Applications to Interview and Micro1 API, checked 20 August 2026.

Interview and assessment formats

• Standard interview: a structured conversation with skills, custom questions and an optional audio or written exercise. It fits repeatable first interviews with clear role requirements.
• Résumé-based interview: Zara uses the candidate’s résumé to tailor the conversation to experience and skills. It may create more context than a common script, but makes quality more dependent on the completeness and accuracy of the résumé.
• Technical interview: a software-engineering format with a coding exercise is documented. It can add practical evidence, but the task quality and accessibility still need employer-led testing.
• Question-only or exercise-only format: custom questions, a coding exercise or a custom exercise can create a narrower, more readily auditable assessment scope.
• Human-data interview: this combines a standard interview with a data-annotation exercise; Micro1 recommends it for selecting experts to train AI models.

Source: Micro1, Types of Interviews, checked 20 August 2026.

Proctoring adds integrity checks and a higher level of intrusion

Optional proctoring requires whole-screen sharing, webcam use throughout the interview and a one-screen setup, according to Micro1. The system logs the environment and flags potentially suspicious behaviour; when proctoring is enforced, the interview is suspended if screen sharing ends or the webcam is turned off. Micro1 states that its Ava model has a 93% cheating-detection success rate. That vendor claim does not by itself establish a false-positive rate, performance for a particular role or suitability for a particular language group. A buyer considering proctoring should therefore agree independent test cases, an appeals route and an accessible alternative. Source: Micro1, What is Proctoring?, checked 20 August 2026.

When Micro1 is a good choice – and when another category fits better

This is a functional assessment based on public documentation, not a statement about a particular contract or implementation. Micro1 is explicitly a good choice for teams building a consistent, asynchronous early assessment that combines conversation, task and subsequent human review. It is not automatically the right choice simply because a team has many applicants: where roles are highly individual, criteria are unclear or EU data residency is mandatory, the organisational and governance work may outweigh the automation benefit.

Pricing, contract scope and business case

No public price, package overview or rate card for Zara or the AI Interviewer was found as of 20 August 2026. The official documentation directs buyers to book a demo. It is therefore not possible to state reliably whether an offer is priced by seats, interviews, candidates, enabled functions, implementation effort or a combination of these. The public sources also do not establish a minimum term, notice period, included support, data export, recording retention, price for proctoring or technical assessments, or the cost of exceeding a purchased volume. Source: Micro1, AI Interviewer, checked 20 August 2026.

The publicly documented functions – AI interviews, reports, candidate pipeline, résumé score, CSV import, ATS import, job-board connection and API – establish what the platform can do technically. They do not establish that every function is included in every commercial tier. The proposal should consequently list inclusions and exclusions line by line: number of jobs and users, completed interviews included and charged separately, coding exercises, proctoring, API and webhook access, ATS implementation, one-off services, support response times, recording storage, export and deletion at contract end.

A useful economics model does not need an invented headcount threshold. Micro1 pays back at a sufficient number of comparable, completed first-round interviews, not at a particular number of recruiters. Use this rule: divide annual software, implementation and integration costs by the expected number of interviews actually completed. Compare that figure with the evidenced time saved per completion multiplied by the fully loaded internal recruiting-hour cost. Explicitly deduct time for re-interviews, technical support, pilot operation, human report review and any alternative pathway needed for candidates. A small team hiring repeatedly into standard roles may therefore have a stronger case than a large team making a few highly bespoke hires; without a price proposal, no credible euro threshold or team-size threshold can be calculated.

What public users and candidates report

Public feedback is mixed and is primarily about candidate or project-participant experience, rather than an independent B2B product comparison. Positive Trustpilot reviewers mention a clear, organised application process, relevant follow-up questions and a conversational flow that builds on previous answers. This is consistent with Zara’s documented adaptive, structured-interview approach, but it is not independent proof of scoring quality or hiring outcomes. Source: Trustpilot, Micro1 reviews, checked 20 August 2026.

Criticism on the same platform includes fast pacing, repeated questions, time pressure, technical interruptions and little or unhelpful feedback after interviews. Some individual reviewers also suspect that interviews are used more for data collection than for a live opportunity. These are individual, independently unverified claims, not established facts. They remain relevant because Micro1’s own privacy notice describes possible use of de-identified audio, video, transcript and assessment material to improve its models. A pilot should therefore measure not just scores but candidate drop-off, complaint reasons, proctoring false alarms, repeat questions, support cases and clarity of communication. Source: Trustpilot, public individual reviews, checked 20 August 2026; Source: Micro1 Candidate Privacy Policy, last updated 21 July 2026.

DACH, EU, GDPR and EU AI Act considerations

For DACH buyers, language should not be treated as a minor configuration detail. Micro1 allows interview-language selection and describes assessment of English, Spanish, French or another supported language. The public pages reviewed do not provide a binding supported-language list or an explicit German-language commitment. Languages also cannot be mixed within an interview, and the documentation says a language change may negatively affect the score. Test German invitations, job vocabulary, regional accents, follow-up questions, accessibility and code-switching before rollout. Source: Micro1, Create Interview and Language Proficiency, checked 20 August 2026.

No public EU-hosting or EU-data-residency commitment was found. The candidate privacy policy says that Micro1 is located in the United States and that information is transferred to and processed in the United States and other locations. It identifies EU standard contractual clauses as a possible transfer safeguard. The policy covers contact, résumé and profile information, audio and video recordings, screen-share recordings, transcripts, scores, assessment results and proctoring data. It also identifies service providers for functions including video processing, foundational models, cloud hosting and data storage, but the page reviewed does not provide a public subprocessors list or a customer-specific data-processing agreement. That is not evidence that such documents are unavailable; they should be requested before procurement. Source: Micro1 Candidate Privacy Policy, last updated 21 July 2026.

Model improvement needs particular attention. Micro1 reserves the right to use de-identified interview information, including audio, video, transcripts and assessment results, to train and enhance proprietary machine-learning models. The policy also says it does not share that data, de-identified or otherwise, with AI model companies to train their models. For an employer, this still does not answer how that processing can be controlled for its own tenant by contract. Ask the relevant privacy and security stakeholders to clarify purpose limitation, opt-out or disabling controls, evidence of de-identification, retention, export, access to recordings and the roles of employer and Micro1. Source: Micro1 Candidate Privacy Policy, last updated 21 July 2026.

Micro1 did not publish specific AI Act conformity or high-risk documentation for Zara in the sources reviewed. Classification depends on the exact use and is not a blanket legal conclusion for every interview. The European Commission identifies AI tools used for employment, worker management and access to self-employment – expressly including CV-sorting software for recruitment – as high-risk use cases. Under the current EU timeline, strict obligations for these systems apply from 2 December 2027, including risk management, high-quality datasets, logging, technical documentation, information for deployers, appropriate human oversight, robustness and accuracy. Zara’s documented human final review is a useful process feature, but it is not evidence of compliance on its own. Source: European Commission, AI Act, checked 20 August 2026.

Eight questions for a vendor evaluation

1. Which exact languages and variants, particularly German and regional accents, are supported and tested with quality measures for our intended workflow?
2. Which inputs affect the résumé score, AI Match Score, language assessment and proctoring score for our role, and how can recruiters inspect, override and document recommendations?
3. Which functions are included and excluded in our proposal: interviews, exercises, proctoring, API, webhooks, ATS integration, storage, support and export?
4. What is the charging unit, minimum term and notice rule, and what do excess volume, implementation, training or additional recording retention cost?
5. Which data are collected, processed and stored for our workflow, which subprocessors are involved, and which EU-transfer safeguards apply?
6. How is possible use of de-identified interview data for model improvement controlled, documented and, if needed, excluded for our tenant?
7. Which data, status values and reports move in both directions with our ATS, and can Micro1 demonstrate that flow in a test environment using our target objects?
8. What evidence and test results can Micro1 provide for bias testing, human oversight, accessibility, data-protection impact assessment and AI Act evaluation?

Frequently asked questions

Does Zara replace recruiters or hiring managers?

No. Micro1’s own documentation positions Zara as support for evaluation while people retain the final decision. In practice, that requires genuine human review with access to criteria, transcript, recording and a documented exception process.

Is Micro1 suitable for technical interviews?

It can be. A software-engineering format and coding exercises are documented. The fit should nevertheless be tested with representative tasks because the presence of a coding module alone does not establish the quality or fairness of an automated assessment.

Does Zara support German?

The documentation supports language choice and refers generally to supported languages, but no binding language list or explicit German product commitment was found. A pilot with realistic candidates, accents and role-specific vocabulary is necessary for German-language hiring.

Is Micro1 GDPR-compliant or hosted in the EU?

No public commitment on those points was found in the sources reviewed. The candidate policy describes processing in the United States and other locations and possible standard contractual clauses; the contractual arrangement for the intended deployment needs review.

What does Zara cost?

Pricing for the AI Interviewer is not public. Obtain a written proposal covering the charging basis, volume bands, implementation, minimum term, add-on functions, storage and exit scenario before calculating a business case.

Is proctoring appropriate for every role?

Not necessarily. It may help where assessment integrity is central, but it requires screen sharing, webcam use and a one-screen setup. Test candidate acceptance, accessibility, false alarms, privacy implications and a fair alternative for candidates who cannot meet those requirements.

Bottom line: Micro1 is a good choice for teams that want structured AI first-round interviews, technical exercises and reviewable candidate reports in one connected workflow. The value case is most plausible for repeatable roles with clear criteria and a meaningful human review step. DACH and EU buyers should validate language quality, data flows, model-improvement use, integration scope, proctoring and every commercial condition in writing and in a pilot before rollout.