First-round screening, done by an agent
The agent interviews every applicant against the same blueprint, scores the answers with a quote behind each conclusion, and puts the whole shortlist on one comparable grid. You read the report, not the CVs.
- 10/10
- applicants interviewed, not just the ones you had time for
- Same blueprint
- for everyone, which is what makes the scores comparable
- Every claim
- backed by a quote from the interview
- Minutes
- from interview finishing to report in your inbox
Built for roles with more applicants than interview slots — operators, warehouse, retail, first-line support.
Why first-round screening breaks
Not because recruiters are careless. Because the format cannot do what is asked of it.
You interview the first eight
Forty applied. Calendar held eight. The other thirty-two were filtered on a CV, which tells you where someone worked, not whether they can do the job.
Every conversation is different
Different questions, different depth, different day. At the end you are comparing impressions, and the impression is mostly of who interviewed well.
Notes are not evidence
A week later nobody can say why candidate four scored worse than candidate two. The reasoning existed; it just was not written down.
Nothing is comparable across time
The person you interviewed in March cannot be put next to the one from June, so good candidates get re-run or quietly lost.
What the agent does
One structured interview per applicant, then the reporting that makes the set usable.
Runs the interview
Voice or text, in the candidate's language, following the blueprint you approved. It asks the prepared follow-ups when an answer is thin.
Scores against your criteria
Ten dimensions by default — clarity, relevance, technical knowledge, problem solving, fit and the rest. Add, remove or reweight any of them.
Shows its evidence
Each score carries the candidate's own words, a confidence level, and the strength of the evidence behind it. Where the interview failed to establish something, it says so rather than guessing.
Checks the knockouts
Shift pattern, licence, right to work, whatever the role requires. Pass, fail or not established — never inferred.
Compares the whole cohort
All candidates on one grid, each cell tinted by score, with the outliers marked. Two to four can be laid over each other on a radar.
Leaves the decision with you
The agent produces evidence and a suggested next step. Who gets hired is a human call, and the report is built to be argued with.
One candidate, as the agent reports them
This is the agent's own output, rendered exactly as it comes out of the product — not a screenshot and not a redesign. The score, the reasoning behind it, the knockout checks, then all ten dimensions with the candidate's own words under each.
The verdict
He showed some of what the role needs — quality control, hand assembly — but the answers often stayed on the surface and he could not give a single concrete example of working with metal parts or advanced tooling. No knockout criterion was breached, so the recommendation is “Maybe”: a middling score and answers that kept thinning out under follow-up.
Quality control and hand assembly in construction, no experience with metal parts — which is the core of this role. Answers were frequently vague and short on examples, which suggests the gap is real rather than a matter of nerves.
Knockouts and must-haves
Health
Said he is in full health and that nothing stands in the way of this kind of work.
"I'm completely healthy."
Getting to work
Has his own car and can also get there by public transport, so the commute is not an obstacle.
"I've got my own car, but sometimes I'd rather take public transport."
Motivation for manual work
Genuinely interested in work where something gets made.
"I enjoy work where I'm making something, so I like working with my hands."
Precision
70 %Says he works on his accuracy and checks the result with measuring instruments.
"I try to improve that precision by getting better and better at the work itself."
Reliability
70 %Worked shifts and organised his time around them.
"I did mornings and afternoons, so I've got a fair bit of experience with that."
Scores across ten dimensions
Mostly clear, but some answers came apart and repeated themselves, so it was hard to follow where he was going.
"Well, I think I'm pretty handy and I enjoy work where I'm making something."
Usually answered what was asked, though sometimes without any specifics, or sidestepped the question.
"It was probably fine, I passed that part on."
Reasonable on hand tools and inspection, but no metal assembly — which is the main thing this role needs.
"I do sort of have experience with hand assembly, but I haven't worked with metal."
Showed no systematic approach; it was most visible on the questions about metal parts.
"I don't know any techniques, I haven't worked with metal parts."
There is potential in hand assembly, but for an engineering assembly role the missing metal experience is a large hole.
Came across as unsure in places, most of all on the metal-assembly questions.
"I already described that, let's move on."
Said little about teamwork or how he adapts, and gave nothing to back the collaboration up.
Answers mostly stopped short of detail and carried no concrete examples.
"Well, exactly what you just said."
Enough to work together, though the answers did contain mistakes.
A large part of the blueprint went uncovered — the answers stayed at the surface and offered no examples.
- A concrete example of metal assembly and of working with advanced tooling.
- Examples of how he solved problems on metal parts.
- Examples of collaboration and adaptability.
- Detailed, specific examples of what he can do.
- Deeper, better-supported answers to the blueprint questions.
Suggested next step
Worth a second round that goes deep on technical ability and on the specifics of what he actually did in previous jobs.
The whole shortlist on one grid
The same ten criteria for every applicant, each cell tinted by its score and the outliers marked, then the top three laid over each other and set side by side. This is what makes a cohort comparable rather than a stack of impressions.
The ranking
You decide the criteria. These ten are the defaults, but any of them can be added, dropped or reweighted — if precision matters more than language in your operation, the blueprint says so, not us. Each cell is tinted by its own score, a star marks a value well above this cohort and a triangle one well below. “Coverage” is how much of the blueprint the interview actually got through.
| Candidate | Score | Recommendation | Clarity of communication | Relevance of answers | Technical knowledge | Problem solving | Fit for the role | Confidence and presence | Behavioural fit | Completeness of answers | Language ability | Depth of answers | Knockouts | Coverage | Date |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Zdenko Choleva choleva@mailinator.com | 48 | Maybe | 6.0 | 7.0★ | 5.0★ | 4.0 | 6.0★ | 4.0 | 3.0 | 5.0 | 6.0 | 5.0★ | ✓ 3 | 0 % | 31. 8. 2026 |
Roman Záskov zaskov@mailinator.com | 46 | Maybe | 6.0 | 5.0 | 4.0 | 3.0 | 5.0 | 5.0 | 5.0★ | 5.0 | 4.0 | 3.0 | ✓ 3 | 33 % | 31. 8. 2026 |
Milan Nedosor nedosor@mailinator.com | 45 | Maybe | 5.0 | 6.0 | 4.0 | 3.0 | 5.0 | 4.0 | 4.0 | 4.0 | 6.0 | 3.0 | ✓ 3 | 73 % | 13. 8. 2026 |
Tomáš Slivka slivka@mailinator.com | 35 | No | 5.0 | 3.0 | 0.0 | 3.0 | 4.0 | 4.0 | 3.0 | 2.0 | 8.0★ | 0.0 | ✕ 1✓ 2 | 0 % | 28. 8. 2026 |
Dalibor Snižanský snizansky@mailinator.com | 29 | No | 4.0 | 3.0 | 4.0 | 3.0 | 2.0 | 4.0 | 3.0 | 3.0 | 6.0 | 3.0 | ✕ 1✓ 2 | 50 % | 28. 8. 2026 |
Ondrej Čierský ciersky@mailinator.com | 25 | No | 4.0 | 5.0 | 2.0 | 2.0 | 3.0 | 4.0 | 2.0 | 3.0 | 6.0 | 2.0 | ✕ 1✓ 2 | 48 % | 12. 8. 2026 |
Juraj Šonkoš sonkos@mailinator.com | 24 | No | 4.0 | 3.0 | 2.0 | 3.0 | 3.0 | 4.0 | 3.0 | 2.0 | 6.0 | 2.0 | ✕ 1✓ 2 | 70 % | 12. 8. 2026 |
Matej Rubovský rubovsky@mailinator.com | 22 | No | 3.0 | 4.0 | 0.0 | 1.0 | 2.0 | 1.0▼ | 2.0 | 1.0 | 5.0 | 1.0 | ✕ 1✓ 2 | 25 % | 28. 8. 2026 |
Martin Ochotnický ochotnicky@mailinator.com | 5 | No | 2.0 | 0.0▼ | 0.0 | 0.0▼ | 0.0▼ | 2.0 | 0.0▼ | 0.0 | 2.0▼ | 0.0 | ✓ 3 | 0 % | 12. 8. 2026 |
Roman Skordeva skordeva@mailinator.com | 0 | No | 1.0▼ | 0.0▼ | 0.0 | 0.0▼ | 0.0▼ | 1.0▼ | 0.0▼ | 0.0 | 1.0▼ | 0.0 | ✕ 2✓ 1 | 10 % | 12. 8. 2026 |
The top three on one chart
Two to four candidates can be drawn at once across all ten areas. It shows immediately that they do not differ overall — they differ in what they are good at.
Three side by side
Exactly what the recruiter sees on the compare screen: strengths, risks, the candidate's own words, and how much of each question area the interview covered.
Zdenko Choleva
- Experience of two-shift operation.
- Knows basic tool maintenance and safety practice.
- Thin deeper technical grasp and problem solving.
- Language barrier in English.
"Erm, I only really like manual work."
Consider a further round probing technical ability and the specifics of previous experience.
Roman Záskov
- Experience of two-shift operation.
- Basic hand-tool skills.
- Little experience assembling metal parts.
- Language barrier in English.
"I like working with my hands, so…"
A second conversation focused on metal assembly, and a check on language ability.
Milan Nedosor
- Experience of tyre quality control.
- Competent with measuring instruments.
- Little experience with metal parts.
- Frequently incomplete and surface-level answers.
"At Kaufland, I worked there half a year."
Ask more about specific assembly techniques and whether they are open to training on metal assembly.
Why it works this way
- •The same blueprint for everyone is what makes the numbers comparable. In a free-form conversation you are comparing impressions, not people.
- •Nobody drops out just because one question went badly — the chart shows exactly where they lose and where they gain.
- •A candidate with a lower total may still be the better hire for your operation. That is why you get every number, not just the ranking.
- •Candidates interviewed months apart can still be compared.
How it runs
- 01
You describe the role
A job description is enough to start. The agent drafts the blueprint — question areas, follow-ups, knockout criteria, scoring weights — and you edit it before anything goes live.
- 02
Candidates interview themselves
They get a link and do it when it suits them, including evenings and weekends. No scheduling, no no-shows costing you a slot.
- 03
Reports arrive as they finish
Each one scored, quoted and checked against the knockouts, in your inbox or your ATS within minutes of the candidate hanging up.
- 04
You interview the shortlist
With the screening done and evidenced, your time goes on the three people worth meeting — and you already know what to probe.
FAQ
- Does the agent decide who gets hired?
- No. It produces scores, evidence and a suggested next step. Every hiring decision stays with your team, and the report is deliberately built so you can disagree with it — every conclusion has the quote it came from.
- What stops it being unfair to some candidates?
- The same blueprint for everyone, scoring tied to evidence rather than impression, and an explicit “not established” where the interview did not get there. It will not infer a knockout it was not told. Where a candidate's language is weaker, that is scored as language ability — it does not quietly drag down their technical score.
- Which languages can it interview in?
- Czech, Slovak, English and others on request. The report comes back in the language you work in, regardless of which language the interview ran in.
- Does it connect to our ATS?
- Yes. Candidates and reports move both ways through the API, so screening lives where your recruiters already work rather than in another tab.
- How long before it is running?
- A first role is usually live within a week of the kickoff call: a day or two to agree the blueprint, the rest on connecting it to how you already hire.
Send us one role
Thirty minutes on a call, then we build the blueprint for a role you are hiring for now and show you what the agent produces on it. Free, and you keep the blueprint either way.