Auto-Screen Job Applicant Resumes with AI and Score Them in a Sheet
An n8n workflow that catches each new job application, extracts the resume text, has Claude score the candidate against your must-have criteria, and logs a ranked row to Google Sheets — so recruiters spend time on the top 20% instead of reading every CV.
- 1
Catch the application
Add a
Webhooknode and point your ATS or application form (Typeform, Greenhouse, a careers page) at it. Extract the candidate name, role and resume file URL with aSetnode. - 2
Extract the resume text
An
HTTP Requestdownloads the resume file, then anExtract From Filenode pulls the plain text. This is what Claude reads — no manual copy-pasting of CVs. - 3
Score with Claude
An
HTTP Requestto Claude sends the resume plus your must-have criteria and asks for strict JSON:score0-100,verdict(advance/maybe/pass), and a one-linereason. Instruct it to ignore name, age, gender and photo. - 4
Log the ranked row
A
Google Sheetsnode appends name, role, score, verdict, reason and a link to the resume. Sort by score and your shortlist builds itself — recruiters open the sheet to the strongest candidates first.
Frequently asked questions
Isn't AI resume screening biased or risky?
It can be if you let it judge protected traits. The prompt here scores only job-relevant, must-have criteria you define (years with a tool, a certification, location eligibility) and is told to ignore name, age, gender and photo. It ranks for a human to review — it never auto-rejects — which keeps a person accountable for every decision.
How does it read PDF resumes?
The application webhook usually includes a file URL. An HTTP Request downloads it and n8n's Extract From File node (or a PDF text step) pulls the text, which is passed to Claude. For plain-text or Google Docs applications you can skip the extraction step.