Creating Custom AI Assistants — Custom GPTs, Gems, Projects & More (AIAA)

AI - Artificial Intelligence, AI for End Users

Every AI chatbot is only as capable as the instructions and data you give it. This practical course shows how to turn a generic chat into a specialist by feeding it knowledge, context and rules, using Custom GPTs, Gems, Projects, Spaces and Agents.

Led by an experienced instructor you will build an assistant from real company data: organize documents and knowledge, craft robust system instructions, test and refine behavior, set sharing and access, and apply practical techniques to get reliable outputs without deep prompt engineering.

Location, current course term



Praha + online (volitelně)
5/15/2026 - 5/15/2026 CZECH
Order 		7/10/2026 - 7/10/2026 CZECH
Order 		9/4/2026 - 9/4/2026 CZECH
Order
Custom Customized Training (date, location, content, duration)

The course:

Hide detail
  • Why build a dedicated AI assistant rather than just use chat
    1. Difference between one-off queries and a tailored assistant
    2. What makes a good assistant: instructions + knowledge + context
    3. When a custom assistant pays off and when a chat is enough
    4. Security and data protection – what AI sees, stores and shares
    5. Overview of platforms and their approaches (Custom GPT, Gem, Project, Space, Agent)
    6. Practical: create a NotebookLM notebook from your materials, prepare sources for the assistant
    7. Practical: write an assistant instruction using both techniques taught
  • Organizing knowledge before you build an assistant
    1. Google NotebookLM – turning your documents into verifiable knowledge
    2. Uploading sources, automatic citations and verifiable answers
    3. Audio overviews – auto-generated podcasts from your documents
    4. ChatGPT Projects – organizing conversations and resources into projects
    5. OneNote with Copilot – using existing notes as an AI source
    6. When to use a notebook and when to build an assistant directly
    7. Practical: create a NotebookLM notebook from your materials, prepare sources for the assistant
  • How to create effective instructions for an AI assistant
    1. Anatomy of a system instruction – role, task, rules, style, limits
    2. "AI as confessor" technique
    3. "System prompt" technique
    4. How to describe tone, format and boundaries so the assistant behaves correctly
    5. Common mistakes in instruction design and how to avoid them
    6. Practical: write an instruction for your assistant using both techniques
    7. How to describe tone, format and boundaries so the assistant behaves correctly
  • Building an AI assistant in practice — from idea to specialist
    1. Common workflow across platforms: naming, instructions, knowledge base, testing
    2. Uploading documents and handling company context (PDF, Word, CSV, slides)
    3. Configuring capabilities — search, image generation, data analysis where supported
    4. Testing and tuning — how to tell the assistant works correctly
    5. Sharing with colleagues and publishing options and limits per platform
    6. Key differences between platforms: strengths and limitations
    7. Practical: each participant builds an assistant for a real company task and tests it
  • Platform comparison and strategy selection
    1. Custom GPT vs Gem vs Project vs Space vs Copilot Agent — clear comparison
    2. Choosing a platform by use case (support, internal KB, content creation, research)
    3. Pricing overview and licensing differences
    4. Recommendations for next steps — how to grow and keep your assistant up to date
Assumed knowledge:
Basic experience with an AI chatbot (ChatGPT, Gemini, Claude or Copilot) is required.
Schedule:
1 day (9:00 AM - 5:00 PM )
Course price:
392.00 € ( 474.32 € incl. 21% VAT)
Language: