VendorDesk
🤖 Powered by Claude AI

Screen resumes
10× faster with AI

Stop reading every resume manually. Paste your JD, upload resumes, and get instant fit scores, skill gap analysis, and ranked shortlists — powered by Claude AI.

How the AI Screening Pipeline Works

VendorDesk's AI screening is a four-step pipeline. Each resume is independently analysed against the job description you provide — the AI doesn't just keyword-match, it understands context: seniority level, technology stack depth, industry domain, and communication quality from the resume text itself.

📋
Paste Job Description
Full JD text
📄
Upload Resumes
PDF / DOCX
🧠
Claude AI Analysis
Deep NLP scoring
🏆
Ranked Shortlist
Export to client
10×
Faster shortlisting
94%
Avg. shortlist accuracy
Resumes per requirement

Step 1 — Write a Strong Job Description

The quality of AI scoring is directly proportional to the detail in your JD. The more specific your description, the more accurately Claude can differentiate a strong fit from a weak one.

A good JD for AI screening includes:

  • Required skills with experience levels (e.g. "5+ years Java, 2+ years Kafka")
  • Domain context (BFSI, e-commerce, healthcare, etc.)
  • Nice-to-have skills separate from must-have
  • Work arrangement (WFO/hybrid, shift times, travel requirements)
💡
Paste the client's original JD directly — don't paraphrase. Claude understands the full context of original job postings better than summarised versions.

Step 2 — Upload Candidate Resumes

Open any Requirement and click Screen with AI. You can either upload resume files directly (PDF or DOCX) or screen candidates already in your pipeline who have resumes attached.

VendorDesk extracts the text from each resume automatically and sends it to Claude AI along with the JD. The analysis typically completes in 5–15 seconds per resume.

⚠️
Scanned image resumes (photos of a printed CV) won't parse well. Ask candidates or vendors to submit text-based PDFs or DOCX files for best results.

Step 3 — Understand the AI Score Card

Each screened candidate gets an AI Score Card with a numerical fit score (0–100) and a breakdown across four dimensions:

RK
Rajesh Kumar
Senior Java Developer • 7 yrs exp
Score: 87
Technical Skills
92%
Experience Fit
85%
Domain Match
80%
Communication
88%

Below the score you'll see a plain-English AI Summary — what the candidate is strong in, what's missing, and a one-line recommendation. This is what you send to your client as a talking point.

Step 4 — Build and Export Your Shortlist

After screening, candidates are sorted by score. Select the ones you want to present to the client and click Export Shortlist. VendorDesk generates a clean PDF or shareable link with the candidate profiles and AI summaries — ready to send to the client without any manual formatting.

🎯
A score of 75+ generally means a strong fit worth presenting. Scores below 60 indicate significant skill gaps — useful to understand before the client call, even if you still submit the candidate.

AI Usage & Limits

AI screening consumes credits from your plan's monthly AI request allowance. Each resume screened = 1 AI request. You can view your current usage under Settings → Billing.

On the Pro plan you get 500 AI requests/month; Enterprise plans have custom limits. If you need more, contact support to increase your cap without changing your plan tier.

Try AI screening today

Screen your first 10 resumes free during your trial. No credit card required.

Start Free Trial → All Guides