8 min read

Understanding Graders

Graders are the core of Grader.io — they evaluate form submissions using AI-powered criteria to score and categorize leads automatically.

How Graders Work

When a submission arrives, the grader processes it through several stages:

  1. Data Extraction — Fields are mapped from the form submission to the grader's input schema
  2. Criteria Evaluation — Each criterion is evaluated against the submission data
  3. Score Calculation — Individual criterion scores are weighted and combined
  4. Categorization — The final score determines the lead category (e.g., Hot, Warm, Cold)
  5. Routing — Based on the category, the lead is routed to the appropriate handler

Grader Types

Rule-Based Graders

Rule-based graders use explicit if/then logic to score submissions:

{
  "criteria": [
    {
      "field": "budget",
      "condition": "greaterThan",
      "value": 10000,
      "score": 25
    },
    {
      "field": "company_size",
      "condition": "between",
      "min": 50,
      "max": 500,
      "score": 20
    }
  ]
}

Best for: Simple, predictable scoring with well-defined thresholds.

AI-Powered Graders

AI graders use natural language processing to evaluate free-text responses:

  • Sentiment analysis
  • Keyword extraction
  • Intent detection
  • Relevance scoring

Best for: Complex evaluations where human judgment is typically required.

Hybrid Graders

Combine rule-based and AI scoring for the best of both worlds. Use rules for objective criteria and AI for subjective ones.

Scoring Weights

Each criterion has a weight that determines its impact on the final score:

CriterionWeightMax Score
Budget30%30
Company Size20%20
Industry Match25%25
Response Quality25%25
Total100%100

Categories

Submissions are categorized based on their final score:

  • 🔥 Hot Lead (80-100) — Immediate follow-up required
  • 🌡️ Warm Lead (50-79) — Schedule a call within 48 hours
  • ❄️ Cold Lead (20-49) — Add to nurture campaign
  • ⛔ Disqualified (0-19) — Does not meet minimum criteria

Best Practices

  1. Start simple — Begin with 3-5 criteria and refine over time
  2. Use real data — Calibrate your scoring with actual past submissions
  3. Review regularly — Check your grader's accuracy weekly and adjust weights
  4. A/B test — Run two graders in parallel to find the optimal configuration

Next: Grader Configuration — Set up advanced scoring rules