Module 1 · Lesson 5 Beginner 1 min read
How AI assistants work
What happens when you type a message to ChatGPT or similar tools.
What happens when you send a message
- 1
You send a prompt
Your question or instruction is sent to the AI system.
- 2
The model processes it
The LLM predicts a helpful response based on its training.
- 3
You receive a reply
The answer streams back, often word by word.
Important to remember
The assistant does not search the live internet by default (unless the tool has a search feature enabled). It generates answers from patterns learned during training. That is why it can be wrong — it is predicting likely text, not looking up verified facts.
Some tools also use retrieval (RAG) to pull in documents or web results before answering, which can make responses more accurate for your specific context.
Examples
- •You ask "Write a professional out-of-office email" and get a draft in seconds
- •You paste meeting notes and ask for action items
- •You request "Explain this like I am new to the topic"
- •A company chatbot answers using internal documents it has been given access to
Key points
- ✓AI assistants use LLMs to predict helpful text responses
- ✓They do not automatically know live facts unless search or documents are connected
- ✓Responses are generated, not looked up from a perfect database
- ✓Tools with retrieval (RAG) can use your documents for better answers
Knowledge check
Why might an AI assistant give an incorrect answer?
Choose the best answer, then check your understanding.