Overview of the LEIA general workflow
LEIA (Learning with an Educational Intelligent Assistant) is a platform for designing interactive simulations where students practice realistic scenarios with AI support. The workflow consists of five main stages:
- Design: The instructor chooses a process/behavior (conversational game, information requirements elicitation, or process requirements elicitation), defines the problem and expected solution, and selects/customizes a persona with specific personality traits.
- Configuration: The instructor defines the scenario (conditioning of LEIA), sets operational parameters (time limits, replicable sessions), and optionally provides an auto-grader.
- Student Access: Students or groups are assigned dedicated sessions. The instructor determines if the activity is a single session or spans multiple sessions.
- Submission: Students upload work products (e.g., UML diagrams in Mermaid syntax) for manual or automated evaluation.
- Review: Instructors download session data in CSV or JSON format, including full conversation transcripts, submissions, and metadata for grading or research.