The problem isn't the tool. Nobody taught you how to talk to it. A practical talk on why prompting is a real skill — and how to build it.

ABSTRACT
Almost everyone I know uses AI. Almost nobody prompts well.

It's not a lack of interest. A prompt feels obvious — until the output isn't what you wanted, and you blame the model. We rarely stop to ask what we actually asked for.

This talk is about why prompting isn't instinctive, even for technical people. Language models don't interpret intent — they respond to what's written. That changes everything.

We'll look at the difference between a vague prompt and one that works, why context isn't optional, and how to write a prompt the way you'd write a brief — because that's essentially what it is.

No LLM theory. No unnecessary jargon. Just patterns you can use tomorrow.

**What you'll take away:**
- Why your current prompts probably have one specific problem (and what it is)
- A prompt structure that consistently gets better results
- Three concrete patterns for technical work, communication, and analysis