Concept build for HR teams
AI Resume Screening with Human Approval
Two AI steps turn a stack of PDF resumes into a ranked, explainable shortlist, and an assistant drafts candidate emails that send only when HR clicks Send.
Verified in the code
- extract, then score
- 2-stepextract, then score
- rubric categories
- 5rubric categories
- fit classes
- 4fit classes
- n8n workflows
- 6n8n workflows
The problem
Recruiters lose hours reading resumes that don’t fit, and screening by gut feel is inconsistent and hard to explain.
What we built
The first AI step extracts a structured profile from each resume; the second scores it against the job on a weighted rubric with written reasoning. Resumes can be uploaded or picked up automatically from a Gmail inbox. An email assistant drafts outreach, but sending lives in a separate workflow that only runs on a human click.
Capabilities
What it does
- Weighted five-part rubric: skills, experience, achievements, education and clarity
- Each candidate gets a score breakdown, strengths, gaps, a recommendation and a fit class
- Resumes arriving as Gmail attachments are saved to Google Drive and mixed with uploaded files
- Unreadable PDFs go to an error log instead of stopping the batch
- AI drafts candidate emails; nothing sends until HR clicks Send, and every email is logged
- Live progress, search, filter by fit and CSV export
Under the hood
Screening pipeline
- 1Job description + resumes
- 2PDF text extraction
- 3AI profile extraction
- 4AI rubric scoring
- 5Ranked shortlist
- 6AI email draft → human Send
Stack
Want something like this for your business?
Book a free call. We’ll look at how your business runs and tell you honestly what’s worth automating, and what isn’t.