You've probably tried this already: you ask an AI for something, and it hands back three confident paragraphs of… fluff. Generic, polite, unusable. Most people conclude the tool is overrated. The real problem is smaller — and completely fixable in the next nine minutes.
🎯 By the end of this lesson you'll be able to
Take a vague request from your real work and rewrite it as a crafted ask that gets a usable result on the first try.
First — meet Claude properly
Claude is an AI assistant built by Anthropic, an AI company founded in 2021 with an unusual center of gravity: safety and reliability research. In practice, Claude has earned its reputation exactly where this course trains you: serious writing, careful analysis, and working through long documents — the "get real work done" end of AI.
You'll find it in four places, all the same brain: the web app at claude.ai, the iOS and Android apps, and a desktop app. Under the hood is a family of models at different power levels — fast light ones to deep heavyweights (names like Haiku, Sonnet, Opus) — a choice you'll learn to make deliberately in Module 6.
🖥️ The interface, guided — tap each numbered dot
A faithful illustration of the Claude web app (as of July 12, 2026 — details evolve). The phone apps carry the same essentials.
claude.ai
＋ New chat
💬 Chats
📁 Projects
Recents
Q2 team email
Supplier compare
Report summary
✳️ How can I help you today?
Write your ask here…
＋🌐Model ▾↑
Tap a dot — this map is all you need to start using Claude today.
✅ Plans, honestly (verified July 12, 2026)
The free plan already includes everything this module needs: chat on all devices, file uploads, web search, content generation, and Projects (up to 5). Paid plans (Pro: $17/mo billed annually, or $20 monthly) add higher usage limits, model choice, and extra products like Claude Code — none required here. Usage limits apply on all plans; hit one, wait a little, continue. Anthropic's official pricing page is always the source of truth.
The one idea that changes everything
Claude cannot read your mind. It reads your words — and fills every gap you leave with a guess. A vague request leaves huge guessing space: who's this for? what tone? what shape? how long? Guess wrong on two of those and the output is useless, no matter how smart the model is.
A crafted ask shrinks the guessing space by stating four things in plain language: what you want, who it's for, what shape, and what matters. Watch it happen:
🔬 Try it: same request, two ways example 1 of 3
Clinic: the three beginner traps
Writing keywords, like a search engine
"marketing plan restaurant" gives the model nothing to work with. Claude is a colleague, not a search box — write a request you could hand to a smart assistant: "Draft a one-page marketing plan for…"
Asking for "the best" without saying best for whom
"What's the best CRM?" Best for a 3-person startup or a bank? State your situation and criteria, and the answer becomes useful instead of a generic list.
Stacking three jobs in one sentence
"Summarize this, translate it, and make a slide" = three tasks fighting for attention. Split them, or say clearly they're sequential steps. One ask, one job.
📌 Pocket card — a crafted ask states:
1. What to produce (the task) · 2. Who will read it (audience) · 3. What shape (format) · 4. What matters (length, tone, red lines). Same effort as a vague request. Radically better output.
🎓 Graded exercise — on your work
Think of a real, vague request you'd actually make this week ("write something about the delay", "help with the presentation"…). Rewrite it below as one clear, complete ask you could hand to Claude. Showing the before → after is even better.
Graded 🤖 against a strict rubric (pass: 70%). Specific feedback either way. Your draft autosaves on this device — nothing gets lost.
Independent course by AI Smart Prompts — not affiliated with or endorsed by Anthropic. Claude is a trademark of Anthropic, PBC. Product facts verified July 12, 2026.
🤖 Lesson Coach AI — grounded in this lesson
Hi! I answer questions about this lesson — concepts, examples, how to apply it to your work. I won't write your graded answer, but I'll hint you toward a better one. What's on your mind?