1 Course-Specific AI Chat

Upload all teaching materials for your course — slides, lecture notes, readings, syllabi, class transcripts and more. The system automatically parses, chunks and embeds the content to build an independent RAG knowledge base for that course.
When students ask questions, the AI first retrieves relevant text passages from the current course knowledge base and constructs answers based exclusively on the retrieved materials. Responses are strictly bounded within the course materials — no cross-course content leaks, no generic built-in knowledge from the base model. Every answer includes source citations with specific document names and page numbers, so students can trace back to the original text and answers remain verifiable.

Instructors can set two independent permissions per file:
1.Visible to students: Students can read the full original document
2.Available to AI: The file content is indexed into the knowledge base and can be cited in answers

A hidden-file mode (instructor-only visibility) is supported for answer keys, internal teaching notes and other reference materials — they can inform AI responses without being accessible to students.

2 AI Assignment Grading & Feedback

For open-ended assignments, case analyses, short-answer questions and other subjective task types common in the social sciences, the platform supports AI auto-grading based on instructor-defined grading rubrics.

Instructors can upload existing assignments and rubrics, or have AI generate assignment questions and grading criteria from course materials, which faculty then edit and approve before release. When grading, the AI scores per criterion individually, quotes original text from the student submission as evidence, and provides targeted strength assessments and improvement suggestions — not just a generic total score and comment.

All AI-generated grading results and scores are only released to students after instructor review, editing and approval. Faculty can adjust scores, revise feedback and add comments, retaining full pedagogical authority at all times. AI only provides drafts and supporting recommendations.

Supports single and batch grading, up to three submission versions per student, and separate marking for late submissions. Once grades and feedback are published, they can be exported to the LMS gradebook.

3 Learning Dialogue & Assessment

Replacing traditional fixed-question quiz formats, instructors can define knowledge point ranges, question structures and dialogue rules, and let AI engage students in progressive interactive conversations for both learning and assessment purposes.

• Learning Mode The AI asks one question at a time per knowledge point. Students respond in open text, and the AI provides feedback grounded in course materials, offering hints and follow-up questions for incorrect or partial answers to guide students toward deeper understanding. Untimed and ungraded for formal scores. Ideal for post-class consolidation, pre-class preparation and exam review.
• Assessment Mode For unit quizzes and learning checks, with attempt limits and time constraints. The AI provides neutral guidance without revealing answers, scores by correctness, and delivers concise overall feedback at the end of the dialogue. Typically 3–5 questions, suitable for quick post-class checks.

Compared to traditional quizzes, dialogue-based assessment makes it far harder for students to look up standard answers using generic AI, greatly reducing cheating risk. It also better measures depth of understanding and reasoning processes, rather than just memorization of final conclusions.

4 Integrated Self-Study Toolkit

For any course material enabled by the instructor, students can generate corresponding AI learning tools in one click and complete self-directed learning all in the same environment, no platform switching required.

• AI Summary: Automatically extracts core arguments and logical frameworks from long documents and chapters, reducing reading time
• AI Mind Map: Structurally presents knowledge points by hierarchy to help students map knowledge systems and logical connections
• Practice Question Generation: Auto-generates case-style practice questions with instant feedback for self-testing
• Real-Time Q&A: Ask the course AI questions at any point and get answers grounded in course materials

All self-study tools are marked as optional independent practice and do not count toward formal grades, encouraging students to learn, review and consolidate at their own pace.

5 Learning Analytics

The platform automatically collects and analyzes all student learning behavior data on the platform, providing instructors with multi-dimensional pedagogical insights.

• Material level: average reading time and conversation turns per document
• Dialogue level: first-attempt accuracy, error rate and average response time for learning dialogues
• Knowledge point level: automatically identifies commonly misunderstood concepts across the class and generates instructional insights
• Student level: task completion progress, engagement and learning performance per student

Instructors can intuitively understand student learning progress through the dashboard, accurately pinpoint teaching difficulties, adjust in-class focus and pacing, and direct teaching effort where it is needed most.

6 Seamless LMS Integration

The platform complies with the LTI 1.3 international education interoperability standard and integrates seamlessly with Canvas, Moodle, Brightspace and other leading learning management systems used in universities worldwide.

No custom integration development is required per institution. Capabilities include:

• Single Sign-On (SSO) jump: Users click the tool link inside the LMS and land directly in AI Agora — no second username or password required
• Identity and permission sync: automatically syncs user identity, course enrollment, roles (instructor/student/TA) and access rights
• Grade passback: assignment scores and assessment results can be synced back to the LMS gradebook
• Course content linkage: can pull existing course materials from the LMS to reduce duplicate uploads

The standardized integration approach is compatible with the IT architecture of most North American universities, secure and controllable, suitable for department- and campus-wide scalable deployment, and significantly reduces integration costs and deployment timelines.