Skip to content
CogniYukti

Interviews

Every vendor will automate your first round. Ask what happens when it is wrong.

Automated screening is the easiest thing in this category to sell and the easiest to regret. A round that rejects people you never see is a round whose mistakes you never find out about — and increasingly, one you may have to answer for.

What we do automate

The volume filter — free, instant, and it names the blocker.

The first pass over a large field does not need an AI call and here it does not make one. One action reads every candidate you hold against the role, sorts them into four bands, and for everybody it set aside it names the requirement that did it — with hard blockers kept apart from preferences you are allowed to override.

That is the honest version of automated screening: it narrows the field, it shows its reasoning, it costs nothing, and it rejects nobody. A human moves people. Always.

Where the line sitsnothing crosses it by itself
  • Read the CV and the roleAutomated
    with the source line on every field
  • Score the whole database against a roleAutomated
    free, instant, no AI call
  • Name the requirement that blocked each personAutomated
    hard blockers kept apart from preferences
  • Flag an application against your own knock-out ruleYour rule, overridable
    recorded with its reason — and overridable
  • Move somebody forwardA person
    always
  • Reject somebodyA person
    always
  • Decide what the interview meantA person
    the findings are evidence, not a verdict
One amber row, and it is a rule you wrote about a notice period you set. Everything that decides a person's outcome is on the violet side.
Illustrative
  • Nothing is auto-rejected

    Not by the sweep, not by the deeper read, not by the interview write-up. No part of this product advances or rejects a person.

  • Knock-out rules are the one exception, and they are yours

    A rule you wrote — a notice period a role cannot wait past — flags an application on arrival, with its reason, and a recruiter can override it. The applicant is recorded, not discarded.

  • Protected attributes are refused

    Search that asks about gender, age, caste, religion or disability is declined rather than answered, and you can read the query before it runs.

  • Blind review

    One switch hides names, emails and current employers across results, for you and for your client.

What we automate differently

Not a robot round — a recorded one.

The alternative to delegating the first human round is making it worth having. An interview that is recorded comes back as findings, each carrying the moment it came from: a claim, a risk, a question that was answered halfway. The interviewer gets read as well as the candidate.

Our view is that a thirty-minute conversation reported properly beats an automated round you cannot interrogate. It is also the harder thing to build, which is why fewer people sell it.

And what is not built

Adaptive AI assessment is ours, and it is not wired to a hiring stage.

CognityHire builds Cognisense — adaptive interviewing with voice and text rounds, structured exercises, simulations, proctoring and calibration. It is a real product and a substantial one.

It is not integrated with a hiring stage here. There are columns on the record that anticipate it and no screen that uses them. If an assessment round is central to your process, treat it as a companion product to evaluate on its own terms, and ask us directly where the hookup stands rather than reading it off this page.

The questions we would ask a vendor

Including of ourselves.

Can we delegate a stage to an AI interview today?

Not from inside the hiring module. Cognisense exists as its own product; the stage-level hookup is not built. That is the plainest way we can put it.

Does anything here reject a candidate automatically?

No. Your own knock-out rules flag an application with a reason, and a recruiter can override the flag. Nothing else advances or rejects anyone.

Do you run coding tests against real test cases?

No — and we want to be precise, because this is the sort of thing that gets assumed. Nothing here executes code. Cognisense presents coding exercises and reads the result rather than running it against a test harness, so neither product gives you a compiler verdict. If executed tests are a requirement, use a tool built for that and bring the result in.

So what do we use for a volume first round?

The free pass, then a recorded human round on the shortlist it produces. For most volume hiring that combination is faster than it sounds, because the free pass runs over everybody you already hold rather than only the people who applied this week.

Is AI screening a differentiator?

No, and we are not going to claim it. Every competitor advertises it. The two things here that are ours are the free pass that names the blocker, and the interview that comes back as evidence.

Book a demo

Bring a role with four hundred applicants.

We will run the free pass on the call, and you can decide whether the round you were about to automate still needs automating.