Before prompts, before technique — this is the ten minutes that gets you comfortable just having a conversation. Everything else builds on this.
The biggest thing that trips people up isn't the technology — it's the mental model. A search engine gives you one shot to phrase things perfectly. Claude is closer to messaging a very well-read colleague: you say something, it responds, and you keep talking until you've got what you need.
A search engine
You don't need the "perfect" query. A rough first message is a completely normal way to start.
A back-and-forth
You can always reply again — "make that shorter," "try a different example" — same as talking to a person.
Something you can break
There's no wrong button to press. Worst case, you start a new conversation and try again.
Only as good as what you tell it
It doesn't know your class, your students, or your school unless you say so in the conversation.
This is a mock-up — no real Claude connection here — but the flow is exactly what you'll see on claude.ai. Pick a message below and watch what a reply actually looks like.
Click any message above to see how a reply comes back.
Drag in a document, and Claude reads it before replying. A Study Design PDF, your own notes, last year's workbook — attach it in the same conversation and say "use this" instead of relying on general knowledge. This matters more than almost anything else for accuracy.