A model only knows what is in the prompt. It has no memory of your project, your audience, or what a good answer looks like to you. Think of it as a brilliant new colleague who is eager to help but starts every task with zero background. The clearer your brief, the less the model has to guess.
The lessons cover why short, vague prompts underperform, how to state the task, context, audience, and format up front, and how to iterate on a prompt by reading the output and fixing what was unclear. A good habit to take away: show your prompt to someone with no context and ask whether they would know what to do. If they would be confused, the model probably will be too.
Watch the lessons in order, then continue on to Roles & System Prompts.
An Anthropic prompt engineer shares five practical tips, starting with the most important one: give clear, detailed task descriptions, then structure the prompt so the model knows what is what.
Video 3minAnthropic's Applied AI team builds a real prompt step by step in a live workshop, showing how to state the task, supply context, and refine instructions through iteration.
Video 24minA quick walkthrough of Google's prompting guide: why successful prompts are longer than most people write, how to give the task enough detail, and how to break big requests into smaller steps.
Video 9minDan Cleary on why "just tell the model what to do" is harder than it sounds, how to define what a good answer looks like, and how to use an LLM to help improve your own prompt.
Video 8min