Nillik vs Manual CV Screening
Most small recruitment teams in Indonesia still screen CVs manually — opening each PDF, reading top to bottom, jotting notes in a spreadsheet. It works, but it scales poorly. This page compares manual screening with Nillik's AI screening on the metrics that actually matter: time, cost, consistency, and the trade-offs you might not be thinking about.
What is Nillik?
Nillik is an AI tool that reads CVs the way a recruiter would, applies a fixed scoring formula (Skills 50%, Experience 35%, Education 15%), and ranks candidates in seconds. Recruiters review the top of the ranked list instead of every CV individually — the ranking surfaces strong matches first.
What is manual CV screening?
Manual screening is the traditional approach: a recruiter or hiring manager opens each CV, reads it, mentally compares it to the job requirements, and decides whether to shortlist. Notes typically end up in a spreadsheet or email thread. Time per CV: 3-8 minutes for an attentive read.
Side-by-side comparison
Numbers assume mid-level role, 50 CVs per posting. Real performance varies with CV quality and requirement clarity. Last reviewed 2026-05-16.
| Criterion | Nillik | Manual CV Screening |
|---|---|---|
| Time per 50 CVs | Under 5 minutes (parallel processing) | 3-7 hours of focused reading |
| Cost per 50 CVs | ~Rp 202,000 at Starter rate (101 credits) | Recruiter time at Rp 50k-100k/hour = Rp 150k-700k in salary cost |
| Consistency | Identical scoring formula every time | Varies — fatigue, mood, time of day affect judgment |
| Unconscious bias | Reduced — scoring blind to name, photo, address (depending on what AI extracts) | Present even with best intentions; well-documented in HR research |
| Detailed reasoning per candidate | Yes — AI returns per-criterion explanation | Depends on recruiter notes; often skipped under time pressure |
| Audit trail | Yes — stored per scoring session with timestamps | Spreadsheet or email; easy to lose |
| Edge cases (unusual CVs) | AI may miss nuance, recommend reprocess or human review | Recruiter can spot unusual signal (e.g., career gap explanation, side projects) |
| Setup | Sign up + upload, no training | Zero setup, but requires training new screeners |
| Scaling beyond one role | Linear — same speed for 5 roles at once | Sub-linear — recruiter context-switches between roles |
| Required infrastructure | Internet + browser | Spreadsheet, email, paper notes |
Choose Nillik if...
- You receive more than 20 CVs per role and screening eats your week
- You want consistent, audit-friendly scoring across recruiters
- You want to reduce unconscious bias in early-stage screening
- Your team is small and pulling recruiters into screening hurts other work
- You hire across multiple roles concurrently
Stick with manual screening if...
- You hire one person every 6+ months — AI tooling isn't worth the workflow change
- Your role requires reading between the lines (creative portfolios, unconventional careers, executive search)
- You don't have budget even for credit-based tools
- Your CVs are very short and screening takes under 1 minute each anyway
- Compliance or hiring policy requires human-only screening at this stage
Frequently asked questions
Will AI screening miss good candidates that a human would catch?
Sometimes. AI applies a fixed formula and may underrate candidates with unusual paths (career switchers, self-taught engineers, portfolio-heavy creatives). The recommended pattern is to use Nillik as a first pass — it surfaces top matches fast — then do a human read of borderline cases (typically scores 40-70) to catch outliers.
Is AI screening fair and legal in Indonesia?
Indonesia has no specific regulation prohibiting AI-assisted CV screening as of 2026-05-16. Best practice is to use AI as a decision-support tool with human review of final shortlists, not as the sole decision-maker. Document your scoring criteria for transparency.
How accurate is Nillik compared to a human recruiter?
Accuracy depends on requirement clarity. Vague requirements (e.g., 'good communication skills') produce noisy scores. Specific requirements (e.g., 'React 16+ experience, e-commerce domain') produce reliable rankings. In our experience, well-defined requirements give Nillik ranking that closely matches what an experienced recruiter would produce — at 50-100x the speed.
Can I see why Nillik scored a candidate a certain way?
Yes. Every scored candidate has a detailed analysis showing the AI's reasoning for each criterion (skills, experience, education). You can audit, override, and reprocess as needed.
Try Nillik free
5 free credits, no credit card. Compare against your manual process on real CVs.