Automating candidate screening means removing manual work from the first round of your hiring funnel — so your recruiter's time is spent on shortlisted candidates, not on reading every resume and making initial phone calls.
Done right, this cuts first-round screening time from 1–2 weeks to 24–48 hours. Here's how to do it.
What "automated candidate screening" actually means
Automated screening covers two stages:
Stage 1 — Resume screening: An AI reads each submitted resume, scores it against your job criteria, and produces a ranked shortlist. No recruiter needs to read a resume at this stage.
Stage 2 — AI phone interview: Shortlisted candidates automatically receive an outbound phone call from an AI voice agent that asks your screening questions, records the conversation, and produces a scored transcript. No recruiter needs to be on the call.
Your recruiter's first touchpoint is the shortlist of candidates who've passed both stages.
Step 1: Write a structured job description
Automation starts with a clear JD. The AI uses it to extract screening criteria, so vague language produces vague scoring. Specifically:
- List required skills explicitly — "3+ years Python" not "strong coding background"
- Set a clear experience range — "3–7 years" gives the AI a threshold, "experienced" does not
- Include must-have certifications or degrees if they're genuine requirements
- Note red flags — e.g., "must have worked in a customer-facing role" or "must be based in Bangalore / open to relocation"
Most AI platforms (including Fawin) let you paste your JD and auto-parse the criteria. Review what the AI extracted before running the campaign.
Step 2: Set your screening criteria and thresholds
Once the AI has parsed your JD, configure:
ATS score threshold — Candidates scoring above this move to the phone interview stage. Start at 60–70 for most roles. Run a test batch of 20–30 resumes and manually check what's being filtered out before committing to a threshold.
Must-have skills — Flag these separately from nice-to-haves. A candidate missing a must-have should score much lower than one missing a nice-to-have.
Red flags — Explicit disqualifiers: too-short tenure, missing required certification, location mismatch. These should drop the score significantly or auto-reject.
Step 3: Set up your AI phone interview
Configure the AI agent for the phone screening stage:
Questions — Write 4–6 structured screening questions. These should cover: role-specific experience, situational scenarios, and any specific deal-breakers you need to verify verbally.
Language — If you're hiring for roles where candidates may be more comfortable in Hindi or Hinglish, configure the agent accordingly. Fawin's agent handles English, Hindi, and Hinglish natively.
Call window — Set the hours during which the AI can make outbound calls. Respect candidate local time — 9:30 AM to 6:30 PM is a reasonable default for Indian candidates.
Retry settings — Configure what happens when a candidate misses the call. Fawin retries up to 2 times with a 24-hour gap and refunds the credit if the call never connects.
Step 4: Import candidates and launch
Upload resumes in bulk (PDF, DOC, DOCX) or integrate with your application source. The AI screens all resumes immediately and queues cleared candidates for the phone interview.
Candidates receive an automated notification (email/SMS) that they'll be contacted for a screening interview.
Step 5: Review the shortlist
After the AI phone interview completes, your recruiter sees:
- Candidate resume + ATS score
- Phone interview transcript
- Overall interview score (0–10)
- AI decision: Cleared / Borderline / Not Cleared
- Strengths and weaknesses extracted from the transcript
The recruiter reviews this packet for each shortlisted candidate and decides who moves to a human interview. This typically takes 5–10 minutes per candidate versus the 30–45 minutes a traditional first-round call takes.
Step 6: Handle edge cases
Candidates who miss all retry attempts — They stay in your pipeline as "attempted." Decide whether to manually follow up or mark as dropped.
Borderline candidates — Don't auto-reject these. Review the transcript personally. The AI flags borderline cases specifically because it's uncertain — human judgment is the right call here.
Overrides — Any AI decision should be overridable. If a recruiter sees something in the transcript the AI missed, they should be able to manually advance a candidate.
What to expect: realistic outcomes
| Metric | Manual screening | Automated screening | |---|---|---| | Time to screen 100 resumes | 5–8 hours | 5–10 minutes | | Time to complete 50 phone screens | 25–40 hours | Runs overnight | | Recruiter hours per hire (screening) | 8–15 hours | 1–2 hours | | Candidate drop-off (missed calls) | High (manual follow-up required) | Low (auto-retry) | | Consistency of scoring | Variable (fatigue, mood) | Consistent |
Common mistakes to avoid
Setting the threshold too high on day one. Start conservative, check what you're filtering out, then tighten.
Skipping the JD review step. If you paste a vague JD and don't review the extracted criteria, you'll get vague scoring. Garbage in, garbage out.
Treating the AI decision as final. AI screening is a first pass, not a verdict. Always allow recruiter overrides.
Not telling candidates. Candidates should know they're receiving an AI screening call. It's both an ethical best practice and increasingly a legal requirement in many jurisdictions.
Ignoring borderline cases. These are your highest-signal cases for calibrating the AI. Review them carefully in the first few campaigns.
Getting started
The fastest way to run your first automated screening campaign:
- Upload your JD to Fawin and review the extracted criteria
- Import your first batch of resumes
- Set your ATS threshold and configure the phone interview questions
- Launch — the AI handles the rest
- Come back in 24–48 hours to review the shortlist
Most teams complete their first campaign setup in under 30 minutes.