Assignment
Design and Build Your First AI Agent
1 Context
In this assignment, your task is to design, implement, and document a simple AI Agent.
You will work in groups of 3–4 students (groups will be assigned in due time).
The focus of this assignment is not on a specific programming language or tool, but on understanding and applying the core concepts of Agentic AI, including:
- agent loops
- tool usage
- memory and retrieval
- orchestration
- governance and guardrails
Your agent should demonstrate autonomous problem-solving behavior: it should take user input, reason using a Large Language Model (LLM), and optionally interact with tools, memory, or external systems.
You may implement your agent using Python, n8n, or a hybrid approach.
You can choose one of the following application scenarios (or propose your own, upon approval):
- 🧑💼 Personal Assistant — Supports users with tasks, appointments, or information management
- 📈 Business Analyst Agent — Analyzes text or data (e.g., reviews, reports, sales summaries)
- 🌍 Information Agent — Uses external APIs (e.g., weather, finance, knowledge sources)
- 🗂️ Knowledge Agent — Implements a Retrieval-Augmented Generation (RAG) setup
- 🎓 Learning Coach — Supports learning through retrieval, explanations, and quizzes
2 Task
Your task is to design, implement, and document an AI agent that:
- Implements a basic agent loop
- receives user input
- calls an LLM
- produces an output for the user
- receives user input
- Includes at least one advanced agent capability
- 🔧 Tool usage
- 🧠 Memory
- 📚 Retrieval / RAG
- 🔧 Tool usage
- Implements explicit orchestration logic
- planning, routing, or delegation of actions
- sequencing or decision logic (static or dynamic)
- planning, routing, or delegation of actions
- Considers governance and guardrails
- constraints, policies, limitations, or monitoring
- technical or prompt-based controls
- constraints, policies, limitations, or monitoring
- Documents the agent using the mandatory documentation template (see below)
3 Agent Documentation Template
All teams must use the provided AI Agent Documentation Template for documenting their agent.
- Format: MS Word (
.docx)
- Length: max. 10 pages (excluding optional appendix)
- Template download: AI Agent Documentation Template
The template structures your documentation along the AI Agent Framework discussed in class, including:
- agent architecture
- orchestration logic
- capabilities
- governance and guardrails
⚠️ Submissions that do not use this template will be considered incomplete.
4 Demo Video
In addition to the written documentation, you must submit a short demo video of your agent.
Requirements:
- Length: max. 5 minutes
- Content:
- brief explanation of the agent’s purpose
- live demonstration of at least one interaction
- visible agent behavior (e.g., tool use, orchestration, memory)
- Format: screen recording with audio narration
For recording your screen and audio, you may use the free version of Loom: https://www.loom.com/. You can then simply add the link to your video in your submission in MS Teams.
Other screen recording tools are also allowed. If the recoding is not online, you can upload the video with your submission in MS Teams.
5 Deliverables
Each team must submit:
- AI Agent Documentation
- completed documentation using the mandatory Word template, (≤ 10 pages)
- submitted as
.docxor exported PDF
- Agent Implementation
- Python notebook, n8n workflow, or hybrid solution
- code or workflow does not need to be embedded in the documentation
- Demo Video
- short screencast (≤ 5 minutes)
- link or uploaded video file
- Optional Technical Appendix
- code snippets
- n8n workflow export
- prompt examples
6 Technical Notes
- Any implementation technology may be used (Python, n8n, etc.)
- The agent does not need to be production-ready
- Focus on clarity, reasoning, and design decisions
- Screenshots and diagrams are encouraged
7 Grading Criteria (100%)
| Aspect | Description | Weight |
|---|---|---|
| Agent design & architecture | Clear structure, alignment with agent framework | 20% |
| Orchestration logic | Planning, routing, delegation, agent loop clarity | 15% |
| Capabilities & functionality | Meaningful use of tools, memory, or retrieval | 20% |
| Governance & guardrails | Awareness of constraints, risks, and controls | 10% |
| Documentation quality | Clear, complete, well-structured use of the template | 15% |
| Demo video | Clear explanation and convincing demonstration | 10% |
| Reflection & discussion | Critical reflection, lessons learned, limitations | 10% |
8 Academic Integrity & Use of AI Tools
- You may use AI tools (e.g., ChatGPT, Copilot, n8n AI Assistant) for ideation, coding, or debugging
- You must document how and where AI tools were used
- All submitted work must be understood and explainable by your team
- Cite any external libraries, tools, or sources
9 Submission
- Team size: 3–4 students
- Submission platform: Microsoft Teams → Assignments
- Deadline: Friday, January 31, 2026, 23:59 CET
What to submit:
- AI Agent Documentation (Word or PDF)
- Demo video (file or link)
- Agent implementation (if separate)