AI Interview Platform for Hiring Teams

What runs today is an async interview against the rubric for that job. The candidate answers, spoken or typed. Questions are generated with the job description and can be previewed before anyone is invited. A round goes out by hand, or on its own when a score clears the bar you set. Below the bar, nothing is sent and nobody is rejected. An outbound phone dialler is not connected.

The interview is an AI interview. We do not dress it up as a recruiter ringing them. We also do not publish a show-rate, a pass-rate, or a time saved. This page will not invent one.

1. 03 · First screen

   The facts, filed before a person sits down.

   What runs today is an async, role-specific AI interview — spoken or typed — graded against that job's rubric. The scene beside this is a sample of the facts a first screen should capture. An outbound phone dialler is not connected.

   • Questions are generated with the job description and can be previewed before anyone is invited
   • A round goes out by hand, or on its own when a score clears the bar you set
   • Below the bar, nothing is sent and nobody is rejected — a person decides
   • Notice period, pay and outcome in the scene are a sample checklist, not a live call recording

After the first interview, further rounds are AI interview, multiple choice, or a human round you run yourself. Questions already asked are not asked again. A human round cannot be automatic. There is no slot-booking network behind it yet. You record the outcome on the pipeline. Camera and window-switch checks are not captured.

Use a human round when the signal you need is a conversation your team must have themselves. Use the async interview when you need comparable answers before you spend the panel. The platform does not decide which of those you value. You do, per role.

1. 04 · Interview rounds

   L1, L2, or a round you design.

   Each round says what it tests, how long the candidate has, and whether it is sent by hand or when a score clears a bar. AI interview and multiple-choice rounds run on the platform. A human round is a conversation you hold — tracked here, not generated.

   • Question source per role: the bank generated for the job, weighted to the skills that round tests
   • Rounds you design: AI interview, multiple choice, or a human round you run yourself
   • Questions already asked are not asked again on the next round
   • A human round cannot be automatic. There is no slot-booking network behind it yet — you record the outcome on the pipeline

The output your team is meant to read is a graded report: the answers, the dimension scores, and a short read of what was strong and what was thin. Until grading finishes, the score stays empty. We do not fill a number in to make the shortlist look finished. Hiring-manager comments stay inside your workspace.

Camera and window-switch checks are not captured. A blank integrity panel is empty. It is not a pass, and it is not a silent rejection. A person still decides. The software will not reject anyone on its own.

1. 06 · Evidence

   See the proof, not just the score.

   A graded round leaves the answers, the per-dimension scores, and a short read of what was strong and what was thin. Until grading finishes, the score stays empty. The flag panel in the scene is a sample: camera and window-switch checks are not captured yet, so an empty session is not a pass.

   • Dimension scores and an overall, written only after the answers exist
   • Strong moments and concerns, grounded in what was actually said
   • Empty means not graded — it is never filled in to look complete
   • A flag, when you have one, never auto-rejects. Camera checks are not captured, so a blank panel is not a pass

2. 07 · Decision & offer

   A written brief before you meet anyone.

   Your team gets a graded report per candidate — every round, every score, every flag — so the first human conversation starts at the interview, not at CV triage.

   • Round-by-round breakdown with scores and reasoning
   • Answers, dimension scores and the written summary, only after grading
   • Hiring-manager comments and @mentions kept internal to your workspace
   • Offer release always takes an explicit human action — it is never automated

Illustrative example of the process, not a performance claim. Actual volumes, timelines and outcomes depend on your role, your market and your selection bar.

Rules on the pipeline listen for a graded application or a graded round, then advance or park. When you save a rule, you see how many current graded applicants already clear its bar. A rule cannot reject terminally, and it cannot release an offer. Offer release is an explicit human action from a permitted role. That is the line to put in a security review, because it is the line the product actually draws.

The interviews sit inside a longer hiring flow: JD structuring, ranking, the pipeline board, and the two commercial models. That flow is AI hiring, and the page with the enquiry form is for employers. This page is only the interview.

Full background verification — DigiLocker, PAN, education certificates, EPFO — is not part of the interview and is not live. Do not write it into a candidate communication as if the brief included it. The brief includes interview evidence. Say that.

A useful brief answers three questions before you open a calendar. Which rubric dimension carried the score, and which one was thin? Are the answers actually there, or is the score still empty because grading has not finished? What did the candidate say, in their words, not in a one-line summary?

The rank on the shortlist is explained in writing for the same reason. “Why this person is first” cites skills, mock evidence, and readiness signals. Ungraded applicants sit below that list, with an invite to the mock rather than a silent discard. If the written reason is vague, the round was vague. Fix the rubric before you blame the candidate or the model.

None of this produces a time-to-hire. A panel that reads the brief and still runs its own hour will be slower than a panel that trusts a number. That is a legitimate choice. The product’s claim is only that the hour can start from evidence instead of from a CV. It is not a claim that your bar will be met more often, or that candidates will accept. Bring one open role if you want to see the brief on your own JD. The form for that conversation is on the employers page.

Role-specific AI interviews are included in the six free months. After that, you either continue on the agency model at 8% of CTC per successful hire, or on a per-interview price that is shared in the onboarding meeting rather than on this page. Inventing a per-interview rupee amount would be a fake rate. Bring one open role to the call if you want the number for your volume.