What a 12-minute AI first-round actually produces

Most pages selling first round interview software show you a dashboard screenshot and a list of adjectives. This one shows you the deliverable.

Vettika runs a 12-minute voice interview with each candidate and hands you back two artifacts: a scored report against your rubric, and a full transcript — one copy for you, one for the candidate. That's the whole product. So instead of describing it, let's walk through an actual report, piece by piece.

Everything below comes from our public sample report. It's demonstration data — a fictional candidate, anonymized, run through the product's seeded "Senior Backend Engineer (Go / Postgres)" campaign. The structure, rubric scoring, and transcript format are exactly what every real interview produces.

The header: one score, one recommendation

The top of the report answers the only question you have at 9am with 40 candidates in the queue: advance or not?

7.8 / 10 — Advance to technical round
Weighted rubric score across 4 criteria. Recommendation threshold for this campaign: ≥ 7.0 advance, 5.0–6.9 review, < 5.0 decline.

Two things matter here. The score is weighted against your rubric — the criteria and weights you set when you created the campaign, not a generic "communication skills" template. And the thresholds are yours too. If your bar for "advance" is 8.0, set it to 8.0. The software applies your judgment consistently across every candidate; it doesn't substitute its own.

The rubric scores: every number comes with a receipt

This is the part that separates a usable report from a vibe. Each criterion gets a score, a weight, and — the part most tools skip — evidence quoted from the transcript. From the sample report:

Systems depth — weight 3 — 8/10
Evidence (from transcript): Described an end-to-end ingest service ("LiveKit events → Go consumer → Neon, ~40k events/hr"); named the first failure mode unprompted (consumer lag during reconnect storms) and the fix (idempotent upserts + backpressure).

Calibration — weight 2 — 6/10
Evidence (from transcript): Pushed back well once (scope of on-call ownership) but overclaimed on Kafka experience; when the agent followed up, admitted it was "mostly NATS, one Kafka migration." Honest on the follow-up, but the initial claim needed the push.

Notice the 6/10. The report doesn't round everything up to make the product look smart. The candidate overclaimed, the interviewer followed up, the candidate corrected themselves — and the score reflects exactly that, with the receipt attached. When you disagree with a score, you can check the evidence and the transcript and overrule it. You're the decision maker; the report is the paper trail.

Strengths and flags: written for the next interviewer

After the scores, the report gives you two short lists. Strengths, from the sample:

• Concrete production examples for every systems question — no hypotheticals.
• Quantifies outcomes without prompting (latency numbers, event volumes).

And — more useful — flags for the human round:

• Probe Kafka/NATS depth directly — initial answer overstated, corrected only on follow-up.
• On-call ownership expectations: candidate wants RFC-to-ship ownership but hedged on pager duty.

This is the handoff most screening processes lose. The recruiter who did the phone screen knows the candidate hedged on pager duty; the hiring manager doing round two never hears about it. Here it's written down, attached to the candidate, before anyone's memory fades.

The transcript: both sides keep it

Every report includes the full transcript. An excerpt from the sample, mid-interview:

AI: You mentioned Kafka earlier — how much of that pipeline was Kafka in production?
Candidate: …Honestly, mostly NATS. I did one Kafka migration at my last job, but day-to-day it was NATS.

That follow-up is the point of doing a live voice interview instead of a keyword filter or a one-way video. The agent noticed the earlier claim, came back to it, and the answer changed. A resume parser scores the word "Kafka." A first-round conversation scores what's behind it.

And the candidate keeps their transcript and sees their own score. We think that's how it should work: if software is going to evaluate someone, they get to see the evaluation. It also keeps the rubric honest — every score you give is a score the candidate can read.

What it doesn't produce

In the spirit of an honest walkthrough, the things you won't find in the report:

• No hire/no-hire verdict. The recommendation is "advance / review / decline" against your thresholds for the first round. The hiring decision stays with humans — where regulations like NYC Local Law 144 are pushing the whole industry anyway.
• No personality inference or video analysis. The score is built from what the candidate said, quoted back to you. Nothing is inferred from tone, face, or background.
• No claims we can't show you. Everything in this post is on the public sample report — structure, scores, transcript format. When we publish real usage numbers, they'll come from attested production instrumentation, not marketing estimates.

How to evaluate first round interview software (ours included)

If you're comparing tools in this category, the report artifact is the fastest way to cut through demos. Ask every vendor — including us — these four questions:

1. Can I see a full sample report before I sign up? Not a screenshot — the actual artifact, scores, evidence, transcript. (Ours is here.)
2. Are scores tied to my rubric or a generic one? If you can't set criteria, weights, and thresholds per role, you're buying someone else's opinion of your candidates.
3. Does every score come with evidence from the transcript? A number without a quote is unauditable — you can't overrule what you can't inspect.
4. What does the candidate get? A first round sets the tone for your whole process. A tool that ghosts candidates with a black-box score is spending your employer brand.

We keep a set of honest comparison pages against the other tools in this category — including what they do better than us and when you should buy them instead.

If you're evaluating multiple AI recruitment tools, our buyer's guide walks through the 7 questions to ask every vendor (including us).

Try the artifact, not the pitch

The fastest way to evaluate this product is the way you'd evaluate a candidate: evidence first.

1. Read the sample scored report — two minutes.
2. Run your first interviews on a real role. First 3 interviews are free, no card, no subscription.
3. Compare the report against your last manual phone-screen notes. Keep whichever is better.