Review queue

Org-wide review kanban at /dashboard/review — sorting, filters, blind review, fraud signals, bulk actions, keyboard shortcuts, and composite scoring.

The review queue at /dashboard/review is the org-wide hiring board. Candidates appear here after completing journey steps (or when already advanced in the pipeline). It complements the per-position pipeline tab with a cross-role view of everything needing attention.

Layout

The queue is a six-column kanban:

Drag a card between columns to advance (except into Needs review). Each card shows candidate name, position title, composite score badge, and trust signals when available.

Click a candidate name to open their journey report (/dashboard/journey-report/[sessionId]/[candidateId]) or candidate profile if no session exists.

Sorting and filters

Position filter

When your org has more than one open position, a Filter by position dropdown appears in the header. Select a single role or All positions to scope the board.

The API returns up to 200 applications, ordered with needs-review completed journeys first, then by completion time.

Implicit sort order

Cards within a column are sorted by:

1. Column priority (needs review → shortlist → … → rejected)
2. Journey completion time (earliest complete first within the same column)

For position-level sorting (reviewer score, AI score, name), use the Pipeline tab on /dashboard/positions/[id].

Blind review (flag: dei_compliance)

When your org enables blind review:

1. Platform grant: dei_compliance
2. Org toggle: Settings → Compliance (/dashboard/settings/compliance) → Enable blind review for recruiters

Effects in the review queue:

• Candidate name is redacted to a neutral identifier
• Email is hidden (candidateEmail: null in API)
• Journey report links still work; redaction applies per role policy

Hiring managers may see full PII depending on org configuration. Blind review is designed to reduce bias during initial screening, not to hide identity from everyone forever.

Fraud and trust signals

Journey trust summary

Each queue card can show trust signals via CandidateTrustSignals — visit patterns, session anomalies, and related integrity hints collected during the journey. These appear as icons/badges on the card without opening the full report.

Careers apply fraud queue

Applications with elevated fraud scores from careers apply appear on a separate surface:

Dashboard → Candidates → Flagged (/dashboard/candidates/flagged)

Use this list alongside the review queue when you suspect synthetic applications or bot traffic. Fraud score is distinct from AI qualification score — a high fraud score does not automatically reject; it flags for human review.

Video integrity analytics

Org-wide video integrity trends live under Analytics → Video integrity (/dashboard/analytics/video-integrity). Use this for pattern detection; individual candidate flags surface on journey reports.

Bulk actions

The org-wide review kanban focuses on per-card advance/reject via drag-and-drop and column primary actions.

For bulk reject, bulk hire, and multi-select stage moves, use the Pipeline tab on the position page (/dashboard/positions/[id]) where checkbox selection is supported.

When rejecting from the position pipeline, optional actions include:

• Release to talent network (flag: talent_network) — candidate enters /dashboard/network if they opt in
• Mark open to opportunities — consent flag for future outreach

Keyboard shortcuts

The review kanban itself is drag-first; there is no dedicated keyboard column navigation on /dashboard/review today.

Composite scoring formula

The score badge on each card reflects the candidate's overall journey score when the journey session is complete, otherwise the application-level score.

Primary score source

overallScore = journeySession.overallScore   // when session status = COMPLETED
            ?? application.score             // fallback

journeySession.overallScore aggregates scorable journey nodes (video, assessment, culture quiz, etc.) according to position and org weight configuration.

Skill + culture blend (flag: culture_fit_scoring)

When culture fit scoring is enabled and a culture weight is configured on the effective profile:

averageScore = round((1 - cultureWeight) × skillScore + cultureWeight × cultureFitScore)

Where:

• skillScore — composite of resume, video, and assessment signals
• cultureFitScore — alignment with org/position culture profile (0–100)
• cultureWeight — org or group override, typically 0–0.5 (default 0.25)

Culture fit is informational — it adjusts ranking but is never an automatic reject gate.

Position-level weights

Per-position configuration controls:

• Resume / qualification weight
• Video response weight
• Assessment category weights
• Minimum score threshold

See Questions & scoring for rubric-level detail.

Tie-breakers

When scores are equal:

1. Most recent journey completion
2. Manual recruiter pin (position pipeline)
3. Application date (newest last in default sort)

Workflow tips

• Process needs review daily — candidates in that column have finished async evidence and are waiting
• Use position filter during high-volume reqs
• Open journey report before shortlisting — composite score is a sort key, not the decision
• Enable blind review before high-volume campus or DEI-sensitive campaigns
• Check flagged applications when careers traffic spikes

Related guides

• Position pipeline
• Culture fit scoring
• Structured interviews — feedback enforcement before advance