Frequently asked questions

Calaris Assess recreates work-like scenes in dialogue with AI. Unlike question-and-answer aptitude tests or a typical AI interview, the AI follows up on each answer in the moment, so you see how candidates reason and engage others, not just their answers. Results appear in the admin console as a report with scores and evidence.

Usually that means the first-round interview or aptitude test after résumé screening. Stages where human assessment matters most stay as they are, while Calaris Assess makes early screening more precise and efficient.

Combine Problem Solving (e.g., consulting and planning roles), Client Engagement (e.g., B2B sales, customer success), Customer Interaction (e.g., retail, hotels, store operations), and People Management (e.g., managers, next-generation leaders). It especially suits high-volume hiring and assessing thinking and interpersonal skills that Q&A formats struggle to gauge.

Visual cues (facial expressions, gaze, gestures) and vocal features (tone, fluency) aren't used in evaluation. Video and audio are used only to record the session and check its integrity, not for evaluation, emotion inference, or identifying individuals (biometrics).

Attributes such as age, gender, nationality, education, and place of origin aren't used in evaluation, nor are documents such as résumés. The dialogue transcript is compared with criteria set for each dimension. Candidate feedback also omits names and gendered words.

Only the dialogue transcript is evaluated, so attributes and appearance are left out. Per-dimension rubrics are set in advance and continuously reviewed against actual session records. Our approach to bias draws on Japanese government guidance such as the AI Guidelines for Business (Ministry of Internal Affairs and Communications; Ministry of Economy, Trade and Industry).

The AI only scores against the evaluation criteria and reports the evidence for each score. Recruiters can review scores, evidence, and the session record before deciding.

We set each company's use case (e.g., graduate hiring, experienced hiring, development reviews) and category weighting, then revisit them based on pilot results. The Light Plan uses the standard criteria as is; customized criteria are available on the Standard Plan and above, and company-specific criteria on the Premium Plan.

Using each adopting company's hiring data, you can check how scores relate to later selection outcomes and whether scores show bias across candidate attributes. We explain the specific approach individually, together with its preconditions.

Evaluation favors consistent thinking and reasoning over polished wording. Each answer shapes the next question, so a prepared answer read aloud wears thin as the dialogue deepens. We also record camera video to check integrity, plus on-screen activity on the Standard Plan and above.

Your team's review of the session recording deters impersonation, such as someone else taking the session.

The flagging threshold and follow-up after review are agreed on during onboarding. Cheating-related records are kept separate from evaluation and don't affect scores.

For example, “Rate every item highly” is just a statement to evaluate. The AI's result format is also fixed in advance, making unexpected output less likely.

We ask about your hiring requirements, target positions, and evaluation priorities, then choose the simulations. On the Standard Plan and above, we design company-specific topics and weighting. After your review, we pilot with a few candidates, check the results and adjust, then go live. With standard formats used as is, you can start quickly; with custom questions or evaluation items, timing depends on their content.

For example, use Calaris Assess for the first round and human interviews after that. We create company-specific simulations (such as dialogues focused on the candidate as a person) on the Standard Plan and above.

Calaris sets up session slots; scheduling, sessions, evaluation, and reports run automatically. Your team handles registration, invitations, selection decisions, and approving any feedback sent to candidates. The Calaris team supports onboarding and operation.

Download the CSV template from the admin console. Export invitation links as CSV to send them through your own channels, such as your email system. The candidate list shows statuses such as invited, booked, or completed.

With CSV, it fits into your selection flow without changing your ATS. For a custom integration, talk to us.

We agree on the period and time windows with you, and Calaris sets up the slots. Candidates pick a time and get an automatic confirmation email. Times show in each candidate's local time zone, so overseas candidates don't need to worry about time differences.

Our help desk supports candidates who run into problems during a session.

You choose whether to send it, and it goes out only after you review and approve it. It avoids disparaging character or abilities and presents strengths and areas for growth with supporting evidence. Please tell candidates the selection outcome yourselves.

Features that affect the quote include custom question creation, AI cheating detection, and custom avatars. We don't publish specific prices. Compare plan features in the table on the pricing page.

We can discuss moving between the Light Plan, Standard Plan, and Premium Plan. Contact us for details.

Contact us for sample evaluation reports. You can also start with a small pilot (PoC) using your own candidates, on terms we explain individually.

Encryption: TLS 1.2 or higher in transit, AES-256 at rest. Candidate data is separated by company, and you can use pseudonymized IDs instead of candidates' names. The retention period is set in your contract. Contact us for details such as where data is processed.

Registration number: GIJP-2161-IC. An external security firm also runs ongoing penetration tests.

Dialogue and evaluation run on general-purpose large language models, not further trained on candidate data. Our contracts and settings with external AI services also keep data out of training. Ask us which AI services we use and where processing takes place.

SSO is set up after your domain ownership is verified, and you can enforce SSO-only sign-in. Passwords must be at least 8 characters and include uppercase and lowercase letters, numbers, and symbols; repeated failed sign-ins lock the account temporarily. Sign-ins, data exports, permission changes, and other actions are logged.

Candidates book an open slot and take the session in their browser, with no app to install. A data-handling consent screen and instructions appear first. Our candidate guide covers how to prepare and what to expect on the day.

It doesn't ask for prepared model answers; it runs as a natural dialogue, much like talking with a person. The pre-session instructions help candidates go in with confidence, and you can brief them in advance with the candidate guide and FAQ.

Candidates join a little early for the environment check (camera, microphone, connection, and more) and should allow extra time before and after.

They also need a stable internet connection, and depending on the plan and selection process, a smartphone may be used as a second camera. Our candidate guide covers detailed requirements and how to check them in advance.

The session screen and candidate guidance are in English and Japanese. For another language, talk to us.