AI Product Building · Module 1 of 10

What is an AI product?

An AI product uses models like LLMs to solve a real user problem.

📖 Key ideas

Start from a user problem, not from "let us add AI".
Common AI features: chat assistants, summarising, search, classification, extraction, writing help.
Most teams build on model APIs (Claude, GPT, Gemini) instead of training models.
AI outputs are probabilistic: the same input can give different answers.
Success means users save time or get better results, measured with data.

💡 Real-life example: Like hiring a very fast but sometimes overconfident assistant: you design the job so their work is checked and useful.

LLMLarge language model

Model APIService you call to use a hosted model

ProbabilisticOutputs can vary for the same input

Use caseSpecific problem the AI solves