# Best Practices

### Best Practices for Assess for Learning

{% hint style="success" %}
**Pro Tip:** For **best results**, define **detailed rubrics**, **model solutions**, **objectives**, and **aligned pedagogy** within the [Learning Journey Data Model (LDM)](https://documentation.globebyte.com/reasoning-ai/learning-journey-model). This maximizes the AI’s ability to provide **targeted** suggestions and mark with consistent reliability.
{% endhint %}

1. **Comprehensive Rubric Design**
   * The AI can only mark effectively if your rubrics are thorough. Provide **clear performance levels**, **detailed descriptions**, and **weights** where applicable.
2. **Model Solutions for Complex Tasks**
   * For essays or open-ended tasks, adding a **Model Solution** gives the AI a reference point to compare the learner’s work against an ideal.
3. **Objective Alignment**
   * Ensure each assessment or module references the **Objectives** it aims to measure. This helps the AI provide more context-based feedback (e.g., “Level 4: Analyzing,” “Level 6: Creating”).
4. **Leverage Pedagogies**
   * If you adopt Bloom’s or SOLO, the AI references those frameworks to classify and respond to learner queries at the right cognitive level.
5. **Keep LDM Objects Updated**
   * Regularly revise **Modules, Rubrics,** and **Objectives** to reflect evolving course content. The AI relies on accurate data to provide correct insights.
6. **Encourage Learner Conversations**
   * Assess for Learning excels when learners actually engage in **two-way** conversations about their performance. Remind them to ask follow-up questions and explore reasons behind the AI’s marking.
7. **Monitor & Validate**
   * While the AI can automate much of the process, instructors should **spot-check** results—especially for borderline or high-stakes scenarios. Over time, the AI “learns” from consistent data and refined rubrics, improving accuracy.

😎  **Enjoy!**

{% hint style="info" %}
Explore [**Tutor for Learning**](https://documentation.globebyte.com/reasoning-ai/tutor-for-learning)**.** Built for **Agentforce,** Tutor is a powerful extension to our **Assess for Learning** agent, designed to enhance the assessment experience by providing the learner with an AI interface into their assessment data.
{% endhint %}