How AI Screenr Works
How AI Screenr works: configure a role in one click, candidates interview async with voice AI, scored reports in 2 minutes. 3 free interviews, no credit card.
Try FreeTrusted by innovative companies
Three steps from job posting to ranked shortlist
The AI Screenr workflow at a glance — no integration, no scheduling.
Configure the role
Paste a job description for one-click AI setup (or build it manually in ~5 minutes). AI Screenr extracts skills, knockouts, rubric weights, and up to 5 structured question blueprints. Edit anything before launch.
Share the interview link
Drop one link into your ATS, email, SMS, or job board. Candidates interview async on any device — no scheduling, no app install, no account creation. Typical duration 15–25 minutes (configurable 5–60).
Review the scored shortlist
Within 2 minutes of each interview, a rubric-backed report lands in your dashboard with a Strong Yes / Yes / Maybe / No recommendation, dimension scores, evidence quotes, and ranked shortlist.
See a live walkthrough in under 5 minutes with 3 free interviews.
Try Free — No Credit CardStep 1 in action — a realistic job configuration
What AI Screenr produces after one-click setup from a pasted job description. Every field is editable before you launch the interview link.
Senior Product Manager (B2B SaaS)
Job Details
Basic information about the position. The AI reads all of this to calibrate questions and evaluate candidates.
Job Title
Senior Product Manager (B2B SaaS)
Job Family
Product
Product roles emphasise prioritization, discovery, and stakeholder reasoning — the AI calibrates follow-ups around judgement and trade-offs rather than execution detail.
Interview Template
Competency-Based Screen
Allows up to 4 follow-ups per question. Pushes on trade-off reasoning and demands specific past examples — surfaces the difference between experienced PMs and PM-adjacent candidates.
Job Description
We're hiring a senior product manager to own a core B2B workflow surface. You will partner with engineering, design, and go-to-market to discover real customer problems, prioritise ruthlessly, and ship outcomes — not outputs.
Normalized Role Brief
Senior product manager with 5+ years of B2B SaaS experience, a track record of owning an end-to-end product area, and the judgement to make scoping calls without a playbook. Writing-first, opinionated, calm under ambiguity.
Concise 2-3 sentence summary the AI uses instead of the full description for question generation.
Skills
Required skills are assessed with dedicated questions. Preferred skills earn bonus credit when demonstrated.
Required Skills
The AI asks targeted questions about each required skill. 3-7 recommended.
Preferred Skills
Nice-to-have skills that help differentiate candidates who both pass the required bar.
Must-Have Competencies
Behavioral/functional capabilities evaluated pass/fail. The AI uses behavioral questions ('Tell me about a time when...').
Defends trade-off decisions with evidence; can articulate what was cut and why, not only what shipped
Runs real customer conversations, not surveys; distinguishes surface complaints from underlying jobs-to-be-done
Moves engineering + design + GTM together without authority; resolves scoping conflicts with written analysis
Levels: Basic = can do with guidance, Intermediate = independent, Advanced = can teach others, Expert = industry-leading.
Knockout Criteria
Automatic disqualifiers. If triggered, candidate receives 'No' recommendation regardless of other scores.
B2B Experience
Fail if: No prior B2B SaaS product experience at any scale
This role requires B2B-specific judgement — buyer ≠ user, procurement cycles, seat economics
Tenure
Fail if: Less than 5 years of product management experience
Senior level — needs to own an area without scaffolding from day one
The AI asks about each criterion during a dedicated screening phase early in the interview.
Custom Interview Questions
Mandatory questions asked in order before general exploration. The AI follows up if answers are vague.
Walk me through the most consequential prioritization call you have made in the last 12 months. What did you cut, what did you keep, and what do you know now that you didn't know then?
Describe a time your discovery work changed the direction of a planned feature. What did you hear, how did you validate it, and what shipped instead?
Tell me about a cross-functional disagreement you resolved. What was the disagreement, what did you write down, and how did the team decide?
Open-ended questions work best. The AI automatically follows up if answers are vague or incomplete.
Question Blueprints
Structured deep-dive questions with pre-written follow-ups ensuring consistent, fair evaluation across all candidates.
B1. Design the first version of a usage-based billing upgrade prompt for our product. Walk me through the user segment, trigger, copy, and success metric.
Knowledge areas to assess:
Pre-written follow-ups:
F1. How would you avoid prompting users who are already in a renewal conversation?
F2. What is your guardrail metric, and at what threshold would you pull the prompt?
F3. How do you decide between in-product prompt vs. email vs. account-manager outreach?
B2. A top-10 customer asks for an enterprise feature that would cost 2 engineer-quarters and would not benefit any other customer. Walk me through how you decide.
Knowledge areas to assess:
Pre-written follow-ups:
F1. How do you communicate the decision to the account team?
F2. What if the customer threatens to churn?
F3. How would your answer change if the feature was easy to generalise later?
Unlike plain questions where the AI invents follow-ups, blueprints ensure every candidate gets the exact same follow-up questions for fair comparison.
Custom Scoring Rubric
Defines how candidates are scored. Each dimension has a weight that determines its impact on the total score.
| Dimension | Weight | Description |
|---|---|---|
| Communication Clarity | 12% | Does the candidate structure answers crisply, use concrete examples, and avoid hedging? |
| Relevance of Answers | 12% | Does the answer directly address the question asked, or does the candidate pivot to safer ground? |
| Technical Knowledge | 18% | Understanding of product-management fundamentals — discovery, prioritization, metrics, lifecycle |
| Problem-Solving | 14% | Ability to reason through ambiguous scenarios, weigh trade-offs, and defend conclusions under pressure |
| Role Fit | 14% | Match with the actual demands of a senior B2B PM role — writing-first, opinionated, comfortable with ambiguity |
| Confidence & Presence | 6% | Steady under follow-up probes; acknowledges gaps without spiraling |
| Behavioral Fit | 10% | Collaboration signals — how the candidate talks about disagreement, conflict, and team credit |
| Completeness of Answers | 14% | Covers the full question, not only the easy half; volunteers caveats and counter-examples |
Default rubric: Communication, Relevance, Technical Knowledge, Problem-Solving, Role Fit, Confidence, Behavioral Fit, Completeness. Auto-adds Language Proficiency and Blueprint Question Depth dimensions when configured.
Interview Settings
Configure duration, language, tone, and additional instructions.
Duration
25 min
Language
English
Template
Competency-Based Screen
Video
Enabled
Tone / Personality
Friendly but structured. The AI keeps the conversation moving and probes politely when answers are vague.
Adjusts the AI's speaking style but never overrides fairness and neutrality rules.
Company Instructions
We are a mid-stage B2B SaaS company with ~80 engineers. Our product is used by RevOps and sales teams. Mention that the role reports to the VP of Product and partners with a Principal Engineer and a Lead Designer.
Injected into the AI's context so it can reference your company naturally and tailor questions to your environment.
Evaluation Notes
Treat written-communication signals (does the candidate structure answers, reference specific metrics, name stakeholders) as positive. Penalise vague generalities and buzzword-heavy answers.
Passed to the scoring engine as additional context when generating scores. Influences how the AI weighs evidence.
Banned Topics / Compliance
Salary negotiation, references, equity package, compensation structure — these are handled separately after the scored round.
The AI already avoids illegal/discriminatory questions by default. Use this for company-specific restrictions.
How AI Screenr filters a 100-candidate funnel
From applications to shortlist: each stage narrows the pipeline using evidence, not gut feel. Numbers below are typical for a mid-volume tech role.
Applications received
All inbound candidates enter the funnel — from job boards, referrals, ATS autoresponders, and direct outreach.
Knockout criteria
Hard filters you define — minimum experience, work authorization, location, salary range, language. Candidates who fail are flagged, not auto-rejected.
Must-have competencies
Pass / fail on the non-negotiable skills for the role (e.g. senior React depth for a Senior React Developer). Evaluated live during the interview.
Language proficiency
Optional CEFR assessment (A1–C2) in the language you specify, with a dedicated interview phase. Skipped if not configured.
Rubric-scored interview
8 default rubric dimensions (customizable) score every answer on a 0–100 scale with evidence quotes, evidence-quality labels (Strong / Moderate / Weak / None), and per-dimension confidence.
Ranked shortlist
Top-scored candidates with 4-point recommendation, executive summary, strengths / risks, notable quotes, and coverage summary. Ready for hiring-manager review.
Step 3 in action — a realistic AI screening report
Exactly what lands in your dashboard within 2 minutes of the candidate saying goodbye. Every score is backed by transcript evidence, an evidence-quality label, and a confidence value.
Alex Morgan
Confidence: 86%
Recommendation Rationale
Strong senior-PM signal. Alex articulated prioritization trade-offs with specifics — exact metrics, named stakeholders, what got cut and why — across multiple examples. Discovery discipline is genuine: the case of reversing a planned feature after three customer conversations was well-structured and included the anti-narrative (what they expected to hear versus what they actually heard). The visible gap is cross-functional leadership under disagreement; Alex defaults to consensus before pushing, which may be fine or may be a risk depending on the team dynamics. Recommended for the hiring-manager round.
Summary
Five-plus years of B2B SaaS product management with clear area ownership. Strong on discovery and prioritization — concrete examples with metrics and trade-offs. Cross-functional leadership is competent but consensus-leaning; the candidate does not push back early when the team direction is wrong. Writing-first instincts are evident from how answers are structured. Calm under follow-up probes.
Knockout Criteria
Over five years of B2B SaaS work across two companies with clear area ownership. Named specific enterprise-deal scenarios.
Seven years of product management experience. Comfortably above the 5-year minimum.
Must-Have Competencies
Multiple concrete examples with named metrics, explicit trade-offs, and what was cut. Defends decisions without hedging.
Real customer-interview examples — not surveys, not analytics. Distinguished jobs-to-be-done from stated feature requests twice without prompting.
Works well across eng + design + GTM but tends to build consensus before pushing hard. Likely fine for collaborative teams; a risk in environments that need a strong product voice.
Scoring Dimensions
Consistently structured answers with explicit context, decision, and outcome. No filler. Named specific metrics and stakeholders without prompting.
“I cut the Phase 2 reporting work because we had 11% adoption on Phase 1 after four weeks — below our 25% bar. The data team wanted to keep going; the tradeoff was between doubling down on adoption vs. shipping a feature that no one was using yet. Anna on analytics and I wrote a one-page decision memo; we killed it.”
Answers hit the specific scenario asked rather than pivoting to safer ground. When pressed, Alex stayed on the question instead of reframing.
“You asked about a prioritization call I regret. The honest answer is the customer-success dashboard — we built it, adoption was high, but the exec team never used the weekly digest we also shipped. I should have killed the digest in week two, not week six.”
Strong fundamentals on discovery, prioritization frameworks, and metric instrumentation. Missed a small point on activation-metric definition vs. retention but self-corrected when probed.
“For the upgrade prompt I'd use activation defined as 'user completed three saved reports in the first week' as the trigger, not time-based. Activation predicts retention in our data — we know from the month-three cohort analysis.”
Thinks in trade-offs. For the top-10 customer scenario, Alex walked through cost-of-distraction, precedent risk, and three alternatives before reaching a recommendation.
“Two engineer-quarters is not the only cost — there's the precedent cost, the integration-debt cost, and the opportunity cost. I'd first explore whether we can deliver 80% of the value as a services engagement; if not, I'd scope it explicitly as a paid, roadmap-committed item with a price that reflects the opportunity cost.”
Writing-first, opinionated, comfortable being specific. Mentioned three decision memos by name. No playbook-thinking — all examples are team-specific.
“For hard decisions I write a one-pager: the decision, the trade-offs, who disagrees and why, and the reversibility. It goes to the team before the meeting. People come in ready to decide, not ready to present.”
Steady under probes. Acknowledged the two regrets without defensiveness. Slight hedging when pushed on the 'disagreement with leadership' scenario.
“I haven't had a full-blown disagreement with a CPO-level stakeholder — most of mine have been with GTM leads. I can give you one of those if that's useful.”
Collaboration signals are positive but consensus-leaning. Several examples showed Alex pushing back only after team direction had already drifted, not earlier.
“I raised the concern at the fourth prioritization meeting. In hindsight the second meeting was where I should have raised it — by then we were three weeks in and the sunk-cost argument was already forming.”
Volunteered caveats and counter-examples without being asked. Answered both halves of compound questions — what went well and what did not.
“The upgrade-prompt experiment succeeded on conversion but the downstream effect was increased support volume from users who upgraded without understanding the entitlement. That was a miss we should have anticipated.”
Blueprint Question Coverage
B1. Usage-based billing upgrade prompt
+ Defined activation with concrete behavioural criteria rather than time
+ Paired conversion metric with a support-volume guardrail
- Did not discuss what threshold would trigger pulling the prompt
B2. Top-10 customer asking for non-generalizable feature
+ Framed cost beyond headcount — precedent + integration debt
+ Proposed scoped, paid alternative before reaching refusal
- Did not address the counterfactual where the feature could be generalised
Interview Coverage
%
Overall Coverage
Strengths
- Writing-first instincts — decision memos are a default tool, not a ceremony
- Prioritization trade-offs backed by named metrics and specific dates
- Strong discovery discipline — distinguishes jobs-to-be-done from stated asks
- Volunteers counter-examples and regrets without being prompted
Risks
- Consensus-leaning under disagreement — pushes back after drift rather than before
- Limited evidence of CPO-level conflict; all examples are GTM-facing
Notable Quotes
“People come in ready to decide, not ready to present.”
“Two engineer-quarters is not the only cost — there's the precedent cost, the integration-debt cost, and the opportunity cost.”
“The honest answer is the customer-success dashboard — we built it, adoption was high, but the exec team never used the weekly digest.”
Suggested Next Step
Advance to a 60-minute hiring-manager round focused on one 'I disagreed with leadership' scenario and one 'scope-cut defence' case study. Probe for the moments where Alex would push back earlier in the process, not only when consensus has already drifted.
AI Screenr turns a job description into a scored shortlist in four steps. This page walks through the AI interview process end-to-end so you know exactly what happens between clicking "create job" and opening the first ranked report.
- Step 1: Configure the role (one click or ~5 minutes manual)
- Step 2: Candidate completes the voice interview async (typically 15–25 minutes)
- Step 3: AI scores and summarises (under 2 minutes per candidate)
- Step 4: You review the ranked shortlist
No ATS integration required. Works with any existing hiring pipeline.
Try the full AI interview process with 3 free interviews →
Step 1 — Configure the Role
You have two paths into a live interview link. Most teams use the AI-generated path.
Option A — One-Click AI Configuration
Paste a job description (internal or public, any length under 10,000 characters). AI Screenr extracts and populates:
- Title, description, role brief, job family, interview template
- Required skills and preferred skills
- Must-have competencies with required levels (basic / intermediate / advanced / expert)
- Knockout criteria (minimum experience, work authorization, salary range, language, anything you wrote into the JD)
- Custom interview questions
- Up to 5 structured question blueprints — each with must-cover topics, follow-up prompts, and strong / weak answer indicators so the AI knows what a great answer sounds like for this specific role
You review the draft, adjust anything that does not match your actual bar, and save. Typical time: 30 seconds to a minute.
Option B — Manual Configuration
Prefer to build from scratch? The form walks you through every field directly. Allow about 5 minutes for a new role, faster once you have a template to clone.
What Gets Configured
Either path produces the same output:
- Core interview questions — typically 6–10 main areas mapped to rubric dimensions.
- Question blueprints — must-cover topics + follow-ups + answer indicators per question.
- Follow-up depth — how hard the AI pushes on shallow answers, configurable per dimension.
- Knockout criteria — flagged, not auto-rejected; you decide the next step.
- Rubric dimensions — 8 default (fully customizable) plus a 9th language-proficiency dimension when the interview is non-English.
- Language + CEFR target — interview language (57 supported) and whether language proficiency is being assessed (A1–C2).
- Interview duration — 5 to 60 minutes, typically 15–25.
- Video recording — optional, opt-in per role.
- Link expiration — how long the link remains active for candidates.
For concrete role examples see React Developer, Backend Developer, and Sales Manager — each page shows a filled-in configuration and sample report.
Step 2 — Candidate Completes the Voice Interview
This is the only step that involves the candidate. Everything upstream is recruiter-side, everything downstream is automated.
What Candidates See
- The link. One URL, sent via your usual channel — ATS auto-reply, email, SMS, job-board message. No account creation, no app install, no scheduling page.
- Consent and mic check. Before recording begins, the candidate sees a consent screen, grants microphone access, and completes a 10-second mic check.
- Greeting. The AI introduces itself, explains the interview flow, confirms the role being interviewed for, and answers "what happens with this recording?" in plain language.
- The conversation. The AI asks questions, listens, and adapts. Strong answers get acknowledged and pushed deeper; shallow answers get follow-up probes. Candidates can ask the AI to repeat a question, take a pause, or ask clarifying questions mid-conversation.
- Close. The AI wraps up, offers the candidate a chance to ask anything, and confirms what happens next.
What Makes the Interview Fair
Every candidate for the same role gets the same rubric. The AI tailors follow-ups to each candidate's specific answers, so no two transcripts are identical, but all transcripts are scored against identical criteria. The candidate cannot tell from the interview alone whether they are doing well or poorly, which reduces performance anxiety and produces more honest signal.
Timing and Completion
- Typical duration: 15–25 minutes (you can configure 5–60 per role).
- Completion rate: 80–90% — significantly higher than one-way recorded video because there is real interaction.
- If the candidate gets disconnected, they can resume from the same link for up to 24 hours; the interview picks up where it left off. Partial interviews are flagged in the report.
For the async-hiring framing of this flow, see our async interview software page.
Step 3 — AI Scores and Produces the Report
Within 2 minutes of the candidate saying goodbye, a structured report is ready in your dashboard. Here is exactly what lands in the report:
Top of the Report
- Overall score — 0–100 weighted aggregate across all rubric dimensions.
- 4-point hiring recommendation — Strong Yes / Yes / Maybe / No.
- Overall confidence — 0.0 to 1.0 score reflecting how much evidence the AI had to work with.
- Executive summary — 2–3 sentence TL;DR for hiring managers who only read the top.
Dimensional Scores
Each rubric dimension shows:
- Score (0–10, then weighted) with a 1–2 sentence rationale
- Evidence-quality label: Strong / Moderate / Weak / None
- Confidence value per dimension (0.0–1.0)
- Evidence snippets — direct transcript quotes that support the score
- Linked questions — which interview question(s) produced the evidence
- Missing evidence notes — what the rubric expected but the transcript did not reveal
Knockout and Must-Have Results
If you defined them in Step 1:
- Knockout results — triggered / assessed flags plus evidence for each criterion.
- Must-have competency results — pass / fail plus evidence per competency.
Summary Blocks
- Strengths — 3–5 bullets of what stood out.
- Risks — 3–5 bullets of what raised concerns.
- Notable quotes — the most interesting lines the AI pulled from the transcript.
- Suggested next step — a recommendation tuned to the score and evidence quality.
- Coverage summary — what fraction of custom questions, competencies, knockouts, and blueprints were actually covered by the candidate's answers.
Transcript and Recording
- Full transcript of the conversation, searchable and time-stamped.
- Audio recording by default; video recording if you enabled it for this role.
Scoring uses the rubric version saved at the time of the interview. If you tune the rubric mid-pipeline, earlier interviews keep their original scores and the new rubric applies from that point forward — you see clean version history instead of silent recomputation.
Every role page has a sample report. Spot-check a few: QA Engineer, DevOps Engineer, Software Engineer, Product Manager.
Step 4 — You Review the Ranked Shortlist
The dashboard ranks candidates by overall score with knockouts surfaced up front. A typical review cycle:
- Scan the ranked list. Top 20% usually deserve a closer look. Knockout-triggered candidates drop to the bottom.
- Open the top reports. Read the 2–3 sentence summary first, then the strengths / risks bullets, then skim the evidence the AI flagged.
- Decide. Advance, reject, or flag for follow-up. Bulk-action the obvious rejects. Keep the strong-Yes tier for the hiring manager.
- Share with hiring managers. One click gives them a share link, a PDF, or a paste-able summary for Slack / email — no AI Screenr account needed.
Recruiter time per candidate drops from 25–45 minutes of live call plus notes to roughly 5 minutes of reviewing a structured report. For concrete math on the hours saved at typical volumes, see replace screening calls.
Timing at a Glance
| Step | Actor | Typical time |
|---|---|---|
| Configure role (AI one-click) | Recruiter | 30–60 seconds |
| Configure role (manual) | Recruiter | ~5 minutes |
| Candidate interview | Candidate | 15–25 minutes (configurable 5–60) |
| AI scoring and report generation | Automated | Under 2 minutes per candidate |
| Review report per candidate | Recruiter | ~5 minutes |
For a 50-candidate funnel, recruiter time drops from roughly 25 hours (live screens + notes) to roughly 5 hours (report review), freeing 20 hours per week for higher-leverage work.
Example Reports by Role
AI Screenr covers every job category — from software engineering to healthcare, retail, construction, and hospitality. Each role page has a sample interview report so you can see exactly what lands in your dashboard. A selection below spanning technical, specialist, and service-sector roles:
| Role | Category |
|---|---|
| Backend Developer | Technology |
| UX Designer | Design |
| Data Analyst | Technology |
| Financial Analyst | Finance |
| Recruiter | HR |
| Paralegal | Legal |
| Real Estate Agent | Real Estate |
| Construction Manager | Construction |
| Production Manager | Manufacturing |
| Veterinarian | Veterinary |
Or browse all 960+ role-specific AI interview guides by category.
Security & Privacy Through the Interview Flow
Each step of the AI interview process has a defined data-handling contract. Consent is captured before any recording starts — candidates see an explicit consent screen covering what is recorded, how it is used, and who can see it. Audio and transcripts are stored in-region (EU hosting is available for GDPR-sensitive pipelines) with configurable retention windows per role, after which data is automatically purged. Candidates can request deletion of their data at any time via a self-service flow, and we publish a Data Processing Agreement on request. On the hiring side, only authenticated users in your workspace can view reports; shared report links can be scoped to expire automatically. For the full security and compliance posture, see the Security, Privacy & Compliance section on the AI interview software page.
Ready to Try?
Start with 3 free interviews, no credit card. Configure your first role in a minute (one click) or 5 minutes (manual) and see your first scored report the same day.
- AI interview software — the category pillar with full capability breakdown and comparisons.
- Pricing — plans, usage-based pricing, and the free trial.
- Automated candidate screening — automation-mechanics angle.
- Replace screening calls — time-savings ROI framing.
- Async interview software — async-first angle.
- AI interviews for IT hiring — industry playbook for software teams.
Frequently Asked Questions
How long does it take to set up AI Screenr?
What does a candidate see in an AI Screenr interview?
How quickly do I get scores after an interview?
Can I change the rubric mid-pipeline?
How do I share AI Screenr reports with hiring managers?
What happens if a candidate goes silent or gets disconnected?
Can candidates retake an AI interview?
How are interviews delivered to candidates?
Does the AI Screenr report sync back to my ATS?
How is the AI interview scored under the hood?
See it work in under 5 minutes
- One-click job setup
- Candidate interview link
- Scored report in 2 minutes
- Ranked shortlist
No credit card required
See how it works — try 3 interviews free
Start with 3 free interviews — no credit card required.
Try Free