SP AI Teaching Assistant

AI Teaching Assistant leverages Claude AI with MCP, enabling SP lecturers to manage Brightspace D2L via natural language commands cutting admin time so educators can focus on teaching.

Comment

Inspiration

Lecturers often spend countless hours planning weekly lessons, ensuring alignment with course objectives and LMS content. We wanted to reduce this burden and create a solution that not only saves time but also promotes consistency and quality across all classes. The rise of agentic AI inspired us to leverage its capabilities for automating repetitive academic tasks.

What it does

Our project uses agentic AI to automatically generate weekly lesson plans for each module by analysing content from the Learning Management System (LMS). It suggests structured activities according to the learning outcomes based on the Bloom's Taxonomy, and assessments tailored to the module’s objectives. This tool will enable lecturers to track important announcements and highlights regarding the module. This ensures lecturers have ready-to-use, standardized plans while still allowing customization.

How we built it

  • Built a web app where lecturers type questions in plain English to manage courses
  • Frontend: React with a clean terminal-style chat interface
  • Backend: Express.js with real-time messaging via Socket.IO
  • Core Intelligence: Claude AI connected to 16 custom tools
  • Tools can search courses, read materials, and generate lesson plans
  • Example: "Create a lesson plan for Week 3"
    • AI fetches relevant course content from SP's Brightspace LMS
    • Analyzes materials (Word docs, PowerPoints, PDFs)
    • Produces a structured teaching plan
  • Real-time streaming responses let users see the AI thinking and working

Challenges we ran into

  • Data Quality: LMS content varies greatly in structure, making it hard to standardize inputs.
  • Pedagogical Alignment: Ensuring AI-generated plans meet academic standards and learning outcomes.
  • User Adoption: Convincing lecturers to trust and adapt to AI-generated plans.
  • Integration Complexity: Connecting with different LMS platforms and handling authentication securely.
  • Cost: Utilizing an LLM requires an API key, which involves associated budgetary resources.

Accomplishments that we're proud of

  • Reduced lesson planning time from hours to minutes.
  • Created a system that promotes consistency across multiple classes.
  • Successfully demonstrated how agentic AI can support educators without replacing their expertise.

What we learned

  • AI can significantly enhance productivity in education when paired with human oversight.
  • Standardization improves quality but flexibility is key for lecturer acceptance.
  • Data preparation and context understanding are critical for meaningful AI outputs.

What's next for SP AI Teaching Assistant

  • Collaboration Features: Enable lecturers to share and co-create lesson plans.
  • Personalization: Allow AI to tailor plans based on student performance data.
  • Analytics Dashboard: Provide insights on engagement and effectiveness of lesson activities.
  • UAT: Increase the scope of User acceptance test (UAT) to gather feedback for refinement.
  • Curriculum Development: Provide suggestions on curriculum updates based on student’s feedback and lecturer’s input.
  • Assessment Creation: Assist in developing learning assessments, including quizzes and student reflection activities.

Built With

Share this project:

Updates