Built an n8n workflow that screens job applications automatically, made specifically to save recruiters time without taking the decision out of their hands:
- 🧠 Scores each applicant against a rubric stored in Airtable — no hardcoded prompts
- 🚫 Bad-format resumes get flagged for a human, not auto-rejected
- 🛡️ Sanitizes resume text against prompt injection before it hits the LLM
- ✅ Every score comes with reasoning + confidence, never just a number
- 👤 AI does the reading, recruiter clicks Approve/Reject — Airtable updates automatically
For recruiters, this means the tedious first pass (reading, matching against criteria, checking availability/skills) is already done by the time you open a candidate — you're reviewing a clean scorecard instead of a raw resume, and you still make every call.
Question for the group: should AI screening tools ever be allowed to auto-reject, or is "AI recommends, human decides" the right default for hiring?