How the Two-Stage Screening Gate Works

3 min read

PlacementFlow screens every candidate in two stages: a short set of non-negotiables asked first (hard knockouts like location and right to work), then a deeper AI screen with scored questions for the candidates who pass. You stay in the loop between the stages — approving leads on real answers, not guesses.

Stage 1: Non-Negotiables First

Before any deep screening happens, the candidate answers your non-negotiables — the questions marked Ask before review in the question builder. These are asked first on every channel: WhatsApp, web chat, and the SMS-linked web form.

Non-negotiables are hard gates:

A failing answer disqualifies the candidate automatically, with the exact question that fired logged as the reason.
A missing or blank answer to a knockout question also disqualifies — the gate fails closed. A candidate can't pass a knockout by skipping it.
The candidate's best contact email is always collected as the final Stage 1 question, so you can actually reach the people who pass.

This ordering means nobody wastes time: candidates who don't meet the basics find out in about a minute, and you never review a full screening from someone who was never eligible.

The Recruiter Review In Between

For candidates arriving from outreach, their Stage 1 answers land in your review queue before deep screening continues. You approve or decline the lead based on what they actually said — current location, right to work — rather than only their sourced profile data.

Stage 2: The Deep AI Screen

Candidates who pass move into the full conversational screen:

Questions are asked one at a time, in a chat format
Each question carries a weight — critical questions influence the score more
If a free-text answer is vague, the AI probes once for specifics (never more than one follow-up per question)
Stage 1 answers are never re-asked

The AI then scores the response 0–100 and attaches a recommendation.

Scores and Thresholds

Score	Recommendation
70 or above	Advance — strong fit, no critical concerns
50–69	Review — worth a look, has open questions
Below 50	Reject — critical mismatches

These are recommendations, not silent decisions. Every scored response appears in your review queue at Candidates → Screening, and you make the call. (Optional auto-advance exists for high scores, but at lower autonomy levels it always creates an approval for you first.)

What the AI Does Here

Asks your screening questions conversationally and extracts structured answers
Probes vague answers once per question for real detail
Scores completed responses against your weights and thresholds
Never overrides you: disqualification overrides, borderline calls, and advancing candidates are always yours to review — and hard-knockout failures are logged with the exact reason so every decision is auditable

When Scoring Fails: The Amber Warning

If the AI can't produce a valid score (an outage, a timeout), you'll see an amber scoring unavailable warning on the response instead of a number. PlacementFlow never fabricates a score — the candidate's raw answers stay available, you can read them and decide manually, or hit Retry AI Scoring. A background job also retries failed scorings automatically.

What Candidates Experience

A friendly chat — on WhatsApp or the web — that takes a few minutes, asks one question at a time, and is upfront about being an AI assistant with human recruiters reviewing every conversation.

Back to Candidate Screening