Declarative knockout rules plus multi-layer detection of AI-generated and dishonest answers. Filters candidates against your hard requirements, flags suspicious responses, blocks fraud — all before a recruiter sees the application.
Most ATS systems either accept everything (so recruiters drown) or filter on keyword soup (so they miss good candidates). HRBlade screens with rule-based gates plus AI-powered fraud and integrity checks — clear pass/fail signals with explanations.
"Must have AWS, must be EU work-authorized, must speak German B2+" — set boolean conditions per requisition, candidates who fail are auto-rejected with a courteous email.
Our anti-LLM classifier flags responses written by ChatGPT, Claude or Gemini. Detection accuracy 94% on our benchmark across English, Spanish and German submissions.
AI-authorship detection on written answers, IP and session metadata recorded with every application, and a full audit trail on every scoring decision. No facial recognition and no voice biometrics: we do not process biometric data.
Scoring prompts instruct the model to disregard name, gender, age, origin and employment gaps, and that instruction is verified by controlled-comparison bias testing before every model change. Every scoring decision is logged and exportable.
AI ranks each CV against the role with a per-criterion breakdown. Not a black-box score — every line item explains what was matched and what was missing.
Every screening decision is logged with the rule that triggered it. Discoverable in case of regulatory review or candidate inquiry under GDPR Article 22.
Per-requisition rules: location, work authorization, languages, must-have skills, salary band. Mix boolean gates with AI-judged signals.
Each application is scored, gated, fraud-checked and routed in under 30 seconds. Pass-the-bar candidates land in your pipeline; clear rejects get a polite no with reasons.
Borderline candidates (close to threshold but missed one criterion) are flagged for recruiter review with the AI's reasoning attached.
Source: Internal benchmark on 8,000 hand-labeled responses, 2025
Source: Pilot deployment benchmark, 2025
Source: Production p95 latency, 2025
Retail, customer support, warehouses — hundreds of applications per role. Screening cuts 70%+ before a recruiter even opens a CV.
Banking, healthcare, defense — candidates must satisfy strict eligibility (clearance, certifications). Knockout rules enforce them automatically with audit trail.
Engineering and design roles using take-homes — detect Copilot-assisted answers, copy-paste from public solutions and outsourced submissions.
Regex misses synonyms ("Postgres" vs "PostgreSQL"), seniority context and adjacent skills. AI screening understands "5 years SaaS backend" as evidence of multiple skills the recruiter cares about, while regex just matches strings.
30+ connectors: LinkedIn, GitHub, Stack Overflow, Indeed, Apollo, Hunter and more. Auto-deduplication.
Autonomous AI agent with 17 hiring intents. Semantic candidate search across your database in 2 seconds.
Closes the loop on your hiring history: AI scores matched against later performance reviews.
The information on this website is provided for general information only and does not constitute an offer or a binding commitment. The binding agreement is the Terms of Use at workspace.hrblade.com/terms; where anything on this site is inconsistent with them, those Terms prevail. Product features, availability and figures shown here may change without notice.
