Home Page - Home Business Online

Marcus heads HR at a company of about three hundred. When his recruiting lead suggested an AI recruiter, he rolled his eyes. He’d read the horror stories about black box screening. His first question wasn’t about features. It was, “Who’s accountable when it gets someone wrong?”

Good question, honestly. Trust is the real hurdle with any AI interviewer, and no brochure fixes that. So here’s a straight look at what The Cognitive’s page claims, what it doesn’t, and what a careful team should still check on their own.

Who Actually Makes the Hiring Decision?

Humans do. The platform calls itself agentic rather than autonomous. An agent runs each stage, from sourcing to outreach to interview to evaluation, then hands back to a person before the next one starts. Your team approves between stages and makes every hiring decision.

For Marcus, that line matters more than any feature. Nothing is rejected automatically. The software handles volume, and people handle judgment. If something goes sideways, there’s a named human in the chain, not a shrug and a vendor email.

What Does an AI Recruiter Do With a Brief?

A recruiter writes the role in plain English. The AI Recruiter searches about 900M public profiles and weighs each person against the full requirements, not just keywords. Each result carries a “Why this match” line, so the reasoning is visible instead of hidden.

Anyone who misses a must have drops below those who meet it, but nobody is removed. That detail helps with fairness, because a person with an unusual path isn’t silently deleted. A recruiter can still scroll down and take a second look.

How Is Scoring Kept Consistent?

Every candidate for a role is assessed on the same competencies and scored against the same rubric. The AI decides each next question live, from the role, the rubric, the resume and previous answers. What’s held constant is what gets measured and how it’s scored.

That removes interviewer variance, fatigue and mood effects, the sort of thing that makes human screening uneven. Say a screener loved the first candidate of the day and was bored by the ninth. The rubric doesn’t care about the time. It just measures.

Can Anyone Check the AI’s Work?

Yes, and this is the part Marcus liked most. Every score is traceable to a quote from the interview. The report includes a 1 to 5 score per criterion, written feedback, a weighted score out of 100, a transcript and a recording. Reviewers can look at what was actually said.

Up to 5 resume claims the AI probed come back marked verified, refuted or unclear. If a claim looks refuted, a human can open the recording and judge for themselves. Evidence beats opinion, and the team isn’t asked to trust a mystery number.

Does the Platform Claim to Be Bias Free?

No. The page says no system should claim to be perfectly free of bias. The Cognitive’s claim is narrower: consistent, tied to evidence, and open to audit, with humans making all final decisions. Honestly, that’s more believable than a promise of perfection.

There’s also a bias and fairness page on the site that covers the controls in more detail. A careful team should read it, ask questions, and test the system with real roles before relying on it. Skepticism isn’t rude here. It’s exactly what a good HR lead should bring.

What Do Candidates Know Going In?

The invitation tells candidates the interview is run by AI. No tricks, no surprise. They choose their own slot from the window the recruiter opens, evenings and weekends included, and no account is needed. The AI is a live, two way video conversation, not a one way recording.

If the employer enables it, rejected candidates get a feedback email with strengths and gaps. That’s a courtesy most companies skip. It doesn’t prove fairness on its own, but it does show respect for the person on the other side of the screen.

Where Does the AI Interviewer Draw Its Limits?

The AI Interviewer screens first rounds in 10 or 20 minute calls. It doesn’t make offers, it doesn’t post jobs, and it doesn’t move stages in your applicant tracking system. A note with the score, summary and report link goes back to the candidate’s record, and that’s all.

Those limits are reassuring. A tool that knows its lane is easier to trust than one that tries to do everything. Marcus would still run a pilot on one role before rolling it out, which sounds like sensible advice for anyone.

How Should a Cautious Team Start?

Use the 7 day free trial, which includes 100 sourcing credits plus 2 live AI interviews. Try one real role. Read the scorecards the way a skeptical reviewer would, comparing them with your own sense of the candidates. Look for places where the evidence doesn’t match the score.

If the trial helps, plans come after. AI interview plans run from $99 to $999 a month, and sourcing plans start at $99 a month. Yearly billing is 17% off. But trust comes before price, and a pilot is the cheapest way to earn it.

Final Thoughts

Trusting an AI recruiter and AI interviewer doesn’t mean switching off your own judgment. It means checking the evidence, keeping humans in charge, and testing before you scale. Marcus didn’t become a fan overnight, but a clear paper trail got him curious, and that’s a decent start.