Skip to main content

Overview

Arbi AI allows you to configure your rubric to evaluate candidates. The rubric (also known as the assessment matrix) is the key component that scores candidates based on your defined criteria. You can set up your rubric either during step 3 of the project builder or later through your project settings.

Getting Started

Automated Rubric Generation

During project creation, Arbi AI can automatically generate a comprehensive rubric based on your provided job description. This feature:
  • Uses a specialized AI model trained for parsing job requirements
  • Creates a complete initial rubric structure with relevant criteria
  • Establishes baseline weightings for each criterion
  • Provides a ready-to-use foundation for immediate candidate evaluation
While you can begin evaluating candidates immediately with the auto-generated rubric, we recommend reviewing and adjusting the criteria to ensure they perfectly align with your specific needs and priorities.

Quick Setup Steps

  1. Access the rubric configuration through either:
    • Step 3 of the project builder when creating a new project
    • Project settings for existing projects
  2. Select “Configure Rubric”
  3. Choose your rubric setup approach:
    • Generate a complete rubric automatically from your job description
    • Add individual criteria and evaluation summaries manually
  4. Adjust criteria and weights as needed
  5. Save and begin evaluations

Customizing Your Rubric

Criterion Scoring Types

Each criterion in your rubric can be configured with one of two scoring methods:
  1. Boolean (True/False)
    • Simple yes/no evaluation
    • Useful for mandatory requirements such as compliance certifications
  2. Rating Scale
    • Customizable scoring levels
    • Detailed requirement definitions per level
    • Flexible scoring ranges
Example of a detailed rating scale criterion: Years of Experience in Role:
  • Extremely Poor (1): Less than 1 year experience
  • Poor (2): 1-2 years experience
  • Average (3): 2-3 years experience
  • Good (4): 3-5 years experience
  • Perfect (5): 5+ years experience with 3+ years in similar role, preferably in startup/fast-growing tech environment
The more specific you make your rating scale definitions, the more accurate and consistent your evaluations will be.

Core Components

Experience & Proficiency

  • Set minimum/maximum years of experience
  • Define required technical skills with proficiency levels
  • Specify industry-specific expertise requirements
  • Add weightage for different experience types

Education & Training

  • Configure minimum education requirements
  • List mandatory certifications or qualifications
  • Set up scoring for additional degrees/certifications
  • Define relevance scores for different fields of study

Custom Criteria

  • Create position-specific evaluation metrics
  • Define scoring scales (1-5, 1-10, etc.)
  • Add weighted scoring for priority skills
  • Include soft skills assessment criteria

Weighting System

Each criterion in your rubric can be assigned a percentage-based weight to reflect its importance in the evaluation process:
  • Higher weights indicate greater significance in the final evaluation
  • Critical requirements should receive higher weightings
  • Adjust weights to align with position priorities
Carefully consider your weight distribution. Criteria with higher weights will have a stronger influence on candidate rankings and matching scores.

AI-Assisted Rubric Refinement

You can quickly modify your rubric using Arbi AI’s intelligent assistance:
  1. Click the “Modify Rubric” button in your rubric settings
  2. Enter a plain-language description of your desired changes (e.g., “Add more emphasis on cloud architecture experience” or “Include criteria for team leadership skills”)
  3. Arbi AI will automatically adjust your rubric while maintaining scoring consistency
This feature helps you rapidly iterate and refine your evaluation criteria without manual reconfiguration.

Best Practices

  • Keep scoring criteria objective, measurable, and highly specific to your role
  • Use consistent scoring scales across similar criteria
  • Review and adjust weights based on position requirements
  • Include both technical and soft skill evaluations
  • Test your rubric with sample candidates before full deployment
  • Break down complex requirements into specific, measurable components
  • Clearly define what constitutes different proficiency levels
  • Align criteria directly with day-to-day job responsibilities
The quality of your AI-powered assessments directly correlates with the specificity of your rubric. Generic criteria may lead to less meaningful evaluations, while detailed, role-specific requirements will result in more accurate candidate matches.