Your Progress
0%
Module 01
🤖 What is Claude
Understand what Claude is, how it reasons, and what it can and can't do — so you can use it realistically.
A Language Model
Claude is an AI trained by Anthropic on vast text data. It predicts the most useful next token — but through that, it reasons, writes, codes, and analyzes at a remarkably high level.
Constitutional AI
Anthropic trains Claude with a "constitution" of principles — to be helpful, harmless, and honest. It's not just filtered output; it's a model that reasons about ethics.
What It Can Do
Write, summarize, code, translate, brainstorm, explain, analyze, extract, roleplay, plan, and much more — across virtually every domain of text and knowledge.
What It Can't Do
Claude can't browse the web in real-time (unless tools are connected), has a knowledge cutoff, can hallucinate facts, and doesn't have persistent memory by default between sessions.
💬 Example — Asking Claude what it is
You: "What are you, exactly? What can I realistically expect from you?" Claude: I'm Claude, an AI assistant made by Anthropic. I'm good at: — Writing, editing, and summarizing text — Coding in most languages — Reasoning through complex problems step by step — Answering questions from my training knowledge (up to a cutoff) I'm not a search engine — I don't browse the web unless you connect me to a tool that allows that. I can also make mistakes, so for high-stakes facts, always verify.
Think of Claude as a brilliant collaborator, not a perfect oracle. The more context and clarity you give it, the better its outputs become. It works best as a thinking partner.
Capability Explorer
📊 Claude's capability for this combo
Select a task and type your domain to see how well Claude handles it...
Knowledge Check
1. Which statement best describes Claude's knowledge limitations?
2. Constitutional AI means Claude…
Module 02
💬 Your First Conversation
Learn prompt basics — how to start a conversation, set the right tone, and give Claude useful context from the very first message.
State Your Goal First
Begin every prompt with what you want. "Write me a blog post about X." "Explain Y to me." "Help me debug this code." Clarity upfront saves iterations.
Set Tone & Role
Tell Claude who it's talking to and what voice to use. "You're a senior developer reviewing my code" or "Explain this like I'm 12" completely changes the output.
Give Context
The more relevant background you share, the better the response. Claude has no access to your situation unless you describe it — assume it knows nothing about your specific case.
Iterate Naturally
Claude remembers the whole conversation thread. Follow up with "make it shorter", "add more examples", or "change the tone to formal" — treat it like a real dialogue.
💬 Bad vs Good First Message
❌ BAD: "write something about coffee" ✅ GOOD: "Write a 150-word Instagram caption for a specialty coffee brand targeting young professionals. Tone: warm, slightly witty. Include a call-to-action to visit our online store. Brand name: BrewCraft." The difference? Specificity. The good version has: → Length (150 words) → Platform (Instagram) → Audience → Tone (warm, witty) → Goal (drive traffic) → Brand name
The "3 Ws" trick: Who is the audience? What is the output? Why does it matter / what's the goal? Answer these 3 in your first message and you'll get dramatically better first drafts.
First Message Builder
🚀 Your First Message to Claude
Fill in the fields above to generate your first message...
Knowledge Check
1. Claude remembers context from earlier in the conversation because…
2. Which element does NOT improve a first message to Claude?
Module 03
🎯 Getting Better Results
Prompt engineering fundamentals: clarity, context, examples, and iteration to unlock Claude's full potential.
Clarity over Cleverness
Plain, direct language beats elaborate phrasing. "List 5 marketing hooks for a fitness app" works better than "Could you perhaps provide some potential hooks that might work for a fitness-related application?"
Show, Don't Just Tell
Giving Claude an example of what you want (few-shot prompting) dramatically increases accuracy. "Like this example: [example]. Now do it for X" is extremely powerful.
Constrain the Output
Specify format, length, structure. "Respond in bullet points", "Keep it under 200 words", "Use a table with 3 columns: Tool | Use Case | Cost" all give you predictable, usable output.
Iterate Strategically
Don't start over — refine. "Great, now make the intro punchier." "Remove the third bullet." "Add a section about pricing." Build on what's working.
📝 Prompt Engineering Framework (CCTF)
C — Context: "I'm a frontend developer building a React SaaS app." C — Command: "Review this component for performance issues." T — Target: "Audience: my senior dev team in a code review." F — Format: "Use bullet points. Severity: Critical / Minor / Suggestion." Combined: "I'm a frontend dev building a React SaaS. Review this component [paste code] for performance issues. Present findings for my senior team using bullet points, labeled Critical / Minor / Suggestion."
💬 Few-Shot Example Prompt
Generate product taglines in this style: Example input: "Ergonomic office chair" Example output: "Work longer. Hurt less." Example input: "Noise-cancelling headphones" Example output: "Your world. Your rules. Your silence." Now generate one for: "AI writing assistant for marketers"
The Vagueness Trap: "Write me a good email" will always produce generic output. Every word of ambiguity in your prompt is a roll of the dice on what Claude assumes you meant. Be specific — you can always relax constraints later.
Advanced Prompt Builder (CCTF Method)
✨ Your Engineered Prompt
Fill the CCTF fields to build a structured prompt...
Knowledge Check
1. "Few-shot prompting" means…
2. When Claude's first output isn't quite right, you should…
Module 04
🖥️ Claude Desktop App
Understand the three modes of Claude: Chat, Cowork, and Code — and know when to use each one.
Chat Interface
The classic conversational UI. Best for quick questions, brainstorming, writing tasks, and iterative dialogue. Available on web, iOS, Android, and desktop.
Cowork (Desktop)
A collaborative workspace for complex, multi-step tasks with files and documents. Claude can see your screen context and work alongside you across apps like Excel and PowerPoint.
Claude Code (Terminal)
A CLI tool for agentic coding workflows. Claude can read your codebase, make changes, run tests, and iterate autonomously. For serious developers.
Extensions (Chrome, Excel, PPT)
Claude embedded in your existing tools. Claude in Chrome browses with you, Claude in Excel analyzes spreadsheets, Claude in PowerPoint creates slides — all in context.
🗺️ Decision Guide: Which Claude mode to use?
Need to... Use... ───────────────────────────────────────────────────── Ask a question / write content → Claude.ai Chat Work with files + complex workflow → Cowork (Desktop) Automate your codebase → Claude Code (CLI) Browse web + research → Claude in Chrome Analyze / generate spreadsheets → Claude in Excel Create slides quickly → Claude in PowerPoint Build integrations + pipelines → API / Claude Code
Pro tip: The Desktop App gives Claude access to your computer's context — it can see what app you're in and what file you're working on. This makes it far more powerful for actual workflows versus browser-only chat.
Mode Recommender — What should I use?
🎯 Recommended Mode
Select your use case to get a recommendation...
Knowledge Check
1. Claude Code (CLI) is best suited for…
2. Which product lets Claude see your screen and work alongside you across apps?
Module 05
📁 Projects & Knowledge
Organize your Claude work into Projects, inject persistent context, and share files so Claude always has the right background.
What Are Projects?
Projects are dedicated workspaces that group conversations around a topic or goal. They maintain persistent context so Claude doesn't start fresh each time.
Project Instructions
Write a system prompt for the project — your company tone, your preferences, background info. This primes Claude for every conversation inside that project automatically.
File Uploads
Upload PDFs, code files, CSVs, images, and more. Claude reads them in full — making it ideal for analyzing reports, reviewing documents, or working with your data.
Knowledge Base
Add documents once to your project's knowledge base. Claude will reference them across all conversations in the project without needing re-uploads every time.
💬 Good Project Instruction Example
Project: "Content Marketing Assistant" System Instruction: "You are a content strategist for TechNova, a B2B SaaS company targeting mid-size e-commerce businesses. Brand voice: Clear, confident, never jargony. We avoid buzzwords. Preferred format: Short paragraphs, use headers, no more than 800 words per blog post unless specified. Our 3 main product pillars: 1. Inventory Automation 2. AI-Powered Analytics 3. Seamless Integrations Always reference at least one pillar when writing content about us."
Projects are your secret weapon for teams. Instead of everyone writing their own context in every prompt, a shared project means all conversations start from the same informed foundation. New team members onboard instantly.
Project Setup Designer
📋 Your Project Instruction
Fill the fields to generate a project system instruction...
Knowledge Check
1. What is the main advantage of Claude Projects over regular conversations?
2. A Project's "system instruction" is…
Module 06
🎨 Artifacts
Understand what Artifacts are, when Claude creates them, and how to use, edit, and export them.
What Are Artifacts?
Artifacts are self-contained outputs Claude creates in a side panel — documents, code files, HTML pages, SVG graphics, React components. They live alongside the conversation.
When Claude Creates Them
Substantial content (500+ words), code meant to run, HTML/React components, structured docs. Claude auto-decides, or you can request: "Put this in an artifact."
Edit & Iterate
Ask Claude to update the artifact directly: "Add a dark mode toggle", "Refactor into a function", "Change the header style." It updates in place — no copy-paste needed.
Export & Share
Copy the content, download the file, or (for React/HTML) preview it live in the browser. Artifacts can be exported as text, code files, or shared links.
💬 Artifact Request Examples
"Create a React component for a pricing table with 3 tiers. Put it in an artifact so I can preview it." "Write a 1,000-word blog post about remote work trends. Format it as a proper document with sections." "Build me an HTML landing page for my freelance design portfolio. Make it dark-themed and use modern CSS." "Generate an SVG diagram showing a REST API request cycle."
Artifacts + Iteration = Magic: Once Claude creates an artifact, keep refining it in the same thread. Say "make the button bigger", "add error handling", or "translate to Spanish". Each update preserves all previous work.
🎨 Artifact Demo — Interactive Preview
// React Pricing Card Component
const PricingCard = ({ plan, price, features }) => (
<div className="card">
<h2>{plan}</h2>
<p className="price">${price}/mo</p>
<ul>{features.map(f => <li key={f}>{f}</li>)}</ul>
<button>Get Started</button>
</div>
);
Knowledge Check
1. An Artifact is best described as…
2. To update an artifact, you should…
Module 07
🔧 Skills & Tools
Extend Claude's reach by connecting external tools via MCP — from Google Drive to Notion to Slack and beyond.
What Are Tools?
Tools let Claude take actions beyond text — browsing the web, executing code, reading files, calling APIs. Each tool extends what Claude can do in a conversation.
MCP (Model Context Protocol)
MCP is a standard for connecting Claude to external services. Any app that builds an MCP server can be plugged into Claude — enabling deep integrations without custom coding.
Web Search
With web search enabled, Claude can look up real-time information, verify facts, and cite sources. Essential for research tasks where fresh data matters.
Popular Connectors
Google Drive, Gmail, Slack, Notion, GitHub, Asana, Jira, Salesforce, and more. Connect once, and Claude can read, write, and update data across your entire stack.
🔧 How MCP Works (Conceptually)
User: "Find all emails about the Q4 report and summarize them."
Without tools: Claude can only say "I don't have access to your email."
With Gmail MCP: Claude → calls Gmail MCP → searches inbox → reads threads
→ summarizes findings → presents results.
The MCP bridge:
[ Claude ] ←→ [ MCP Server ] ←→ [ External Service (Gmail, Notion, etc.) ]
You connect MCP servers once in Settings → Connectors.
Claude then decides when to call them based on your request.
MCP is the future of AI workflows. Rather than exporting data and pasting it into Claude, MCP lets Claude live inside your actual tools. This transforms Claude from a chatbot into a genuine work automation layer.
🔌 Tool Connector Demo
Click tools to "connect" them and see what Claude can do with each:
Gmail
Emails
Notion
Notes & DB
Google Drive
Files
Slack
Messaging
GitHub
Code repos
Calendar
Events
Knowledge Check
1. What does MCP stand for in the context of Claude?
2. With Gmail MCP connected, Claude can…
Module 08
🔬 Enterprise Search & Research Mode
Use Claude's deep research capabilities to synthesize information from multiple sources with citations and verified grounding.
Deep Research
Research Mode lets Claude run extended searches, synthesize multiple sources, and produce long-form research reports — far beyond what a single web search can do.
Search Grounding
When web search is enabled, Claude grounds its answers in real-time data. Claims are backed by sources, not just training data — critical for time-sensitive facts.
Citing Sources
Claude can cite its sources inline when using search. Ask "cite your sources" or "add footnotes" and it will link back to the web pages it used to build the answer.
Enterprise Search
In enterprise settings, Claude can search across your internal knowledge bases, connected tools (Drive, Notion, Slack), and the web simultaneously to give unified answers.
💬 Research Mode Prompt Examples
"Research the current state of edge AI inference in 2025. Cite your sources. I want a structured report with: 1. Market overview (with numbers) 2. Key players and their approaches 3. Technical challenges remaining 4. 6-month outlook Target length: 600-800 words." "Compare the top 3 no-code app builders available today. Use web search to get current pricing. Present as a comparison table." "Find recent academic papers on transformer attention optimization published in 2024-2025 and summarize the main findings."
Always verify critical facts: Even with web search, Claude can misread or misattribute sources. For high-stakes decisions (medical, legal, financial), treat Claude's research as a starting point — not a final answer.
Research Prompt Builder
🔬 Your Research Prompt
Fill the fields to build a deep research prompt...
Knowledge Check
1. What does "search grounding" mean in Claude's context?
2. For a high-stakes medical decision, Claude's research output should be used as…
Module 09
👥 Use-Cases by Role
Real examples of how developers, writers, analysts, educators, and managers use Claude in their daily work.
Select your role below to see tailored use-cases and example prompts. Every role unlocks different superpowers with Claude.
Role Explorer
💻 Developer
✍️ Writer
📊 Analyst
🎓 Educator
📋 Manager
Code review: paste a function, ask for security and performance issues
Debugging: share the error + code, get root cause analysis
Documentation: "Write JSDoc for this function" or generate a README
Architecture: "Design a microservices architecture for X requirement"
Test writing: "Write Jest unit tests for this module with edge cases"
💬 Developer Prompt
"Review this Python function for performance and security issues. Point out: time complexity, any SQL injection risks, and suggest improvements. Use severity labels."
First drafts: "Write a 500-word article on X in a conversational tone"
Editing: "Improve the clarity and flow of this draft without changing the voice"
SEO: "Rewrite for SEO, targeting keyword 'remote work tools'"
Multiple formats: turn one blog post into social posts, newsletter, and scripts
Research assistance: "What are 10 compelling angles for a story about urban farming?"
💬 Writer Prompt
"Take this rough draft and restructure it to have a stronger hook, cleaner paragraph transitions, and a memorable closing line. Preserve my voice — don't make it sound corporate."
Data interpretation: paste CSV data and ask for trends and anomalies
Report writing: "Turn these raw metrics into an executive summary"
Formula help: "Write an Excel formula to calculate rolling 7-day average"
Visualization advice: "What chart type best shows this kind of data?"
Benchmarking: "Compare our Q3 metrics to industry averages for SaaS"
💬 Analyst Prompt
"Here is our monthly retention data [paste data]. Identify the 3 biggest trends, explain possible causes for each, and suggest 2 actionable experiments to test your hypotheses."
Lesson plans: "Create a 45-min lesson plan on photosynthesis for Grade 8"
Quiz generation: "Write 10 multiple-choice questions with answers on WWII"
Explanations: "Explain quantum entanglement like I'm 14 years old"
Feedback: "Give constructive feedback on this student essay using a rubric"
Differentiation: "Adapt this material for advanced learners and for struggling learners"
💬 Educator Prompt
"Create a lesson plan for a 60-minute session teaching basic Python loops to high school students with no prior coding experience. Include: learning objectives, activities, a coding exercise, and an exit ticket question."
Meeting prep: "Summarize these 5 project updates into a 5-min briefing"
Performance reviews: "Help me write balanced, constructive feedback for this employee"
Strategy docs: "Create a one-pager for our Q1 OKRs based on this outline"
Communication: "Draft a message to the team about the product deadline shift"
Decision-making: "Pros/cons for building vs buying our analytics solution"
💬 Manager Prompt
"Write a 1:1 agenda for a performance check-in with a mid-level engineer who has been struggling with deadlines. Include: 3 open-ended questions, a section to celebrate wins, and a space to co-create an improvement plan."
Knowledge Check
1. A developer asks Claude to "write Jest unit tests for this module with edge cases." This is an example of Claude being used for…
2. For a manager, one of Claude's most powerful use-cases is…
Module 10
📐 Prompting Patterns
Reusable prompt templates you can copy and adapt: summarize, rewrite, extract, brainstorm, code review, and more.
Prompt patterns are reusable recipes. Master these 6 patterns and you can handle 80% of professional use-cases with Claude. Click each to expand the template.
📋 Summarize Pattern Content
▼
💬 Template
Summarize the following [CONTENT_TYPE] for [AUDIENCE]. Format: [bullet points / paragraph / TL;DR / executive summary]. Length: [SHORT=3 bullets | MEDIUM=100 words | LONG=300 words]. Focus on: [key decisions / action items / main arguments / data]. [PASTE CONTENT HERE]
✍️ Rewrite Pattern Writing
▼
💬 Template
Rewrite the following text for [AUDIENCE] in a [TONE] voice. Goals: [improve clarity / fix grammar / make more persuasive / simplify]. Constraints: [keep it under X words / preserve the key message / don't change facts]. Style reference: [example style or "match the existing paragraph structure"]. Original text: [PASTE TEXT HERE]
🔍 Extract Pattern Analysis
▼
💬 Template
From the following [DOCUMENT_TYPE], extract all [ENTITY_TYPE]. Entity types to find: [names / dates / prices / action items / requirements / risks]. Output format: [JSON / table / numbered list / structured markdown]. If something is ambiguous, flag it with a ⚠️ symbol. [PASTE DOCUMENT HERE]
🧠 Brainstorm Pattern Ideation
▼
💬 Template
Generate [NUMBER] [IDEA_TYPE] for [GOAL/CONTEXT]. Target audience: [AUDIENCE]. Constraints: [budget / time / technical / brand constraints]. Style: [realistic / creative / outside-the-box / data-driven]. For each idea, include: [1-line description / why it works / potential challenge]. Avoid: [things to explicitly exclude].
💻 Code Review Pattern Dev
▼
💬 Template
Review the following [LANGUAGE] code. Check for: - [Security vulnerabilities / SQL injection / XSS] - [Performance: time complexity, unnecessary re-renders] - [Code quality: readability, naming, DRY violations] - [Edge cases not handled] Label each finding: 🔴 Critical | 🟡 Minor | 🟢 Suggestion At the end, give an overall score out of 10 and a top-3 priority fixes list. ```[LANGUAGE] [PASTE CODE HERE] ```
📊 Comparison Pattern Analysis
▼
💬 Template
Compare [OPTION_A] vs [OPTION_B] (vs [OPTION_C]) for [USE_CASE]. Evaluation criteria (weight each if needed): - [Cost / Scalability / Ease of use / Community support / Performance] Output format: markdown comparison table. Then: 1 paragraph recommendation for [my specific context: DESCRIBE IT]. Tone: objective, data-driven, no fluff.
Pattern Customizer
📐 Customized Pattern Prompt
Select a pattern and fill in your topic...
Knowledge Check
1. In the Extract Pattern, what does flagging ambiguous items with ⚠️ achieve?
2. The main benefit of using prompt templates/patterns is…
📋 HANDS-ON TASKS
Apply What You've Learned
4 tasks covering all 10 modules. Each task unlocks after the previous one passes. Progress saved automatically.
Task 01 · Modules 1–3
Write 3 Real-World Prompts
Apply what you learned about prompt basics and prompt engineering
Scenario: You're onboarding a new team member. Write 3 prompts for Claude to help with 3 different professional scenarios — one for writing/content, one for analysis/data, and one for code/technical work. Each prompt must demonstrate clarity, context, and format constraints.
Requirements Checklist
Has at least 3 prompts (separated by blank lines or numbers)
At least one prompt mentions a specific audience or role
At least one prompt specifies output format (table / bullet / words)
At least one prompt includes a tone directive (formal / casual / etc.)
Total length is at least 100 characters
Task 02 · Modules 4–5
Design a Project Setup
Create a complete Claude Project configuration for a fictional team
🔒 Complete Task 1 first to unlock
Scenario: You're setting up Claude for a 5-person marketing team at a startup called "Lumeo" — an AI-powered video editing tool. Design their Project setup: name it, write a system instruction, list what files to add to the knowledge base, and describe which Claude mode/product each team member should use.
Requirements Checklist
Includes a project name
Has a system instruction with tone/voice guidelines
Mentions at least 2 specific files for the knowledge base
References at least one Claude mode (Chat / Cowork / Code / extension)
Total length is at least 120 characters
Task 03 · Modules 6–8
Build a Prompt Template with Variables
Create a reusable template using [VARIABLE] placeholders
🔒 Complete Task 2 first to unlock
Scenario: Your team uses Claude for research every week. Build a reusable prompt template for research reports that includes variable placeholders like [TOPIC], [AUDIENCE], [FORMAT], and [DEPTH]. The template should be usable by any team member for any research need.
Requirements Checklist
Has at least 3 [VARIABLE] placeholders in brackets
Includes [TOPIC] or [SUBJECT] variable
Specifies an output format variable or constraint
Includes instructions for depth, length, or audience
Total length is at least 100 characters
Task 04 · All Modules
Design a Full Claude Workflow
Create a complete end-to-end Claude workflow for a role of your choice
🔒 Complete Task 3 first to unlock
Scenario: You're writing a "Claude Playbook" for your role (developer, writer, analyst, educator, manager — your choice). Describe: 1) Your role and top 3 Claude use-cases, 2) What Projects and knowledge base you'd set up, 3) Which tools/MCP you'd connect, 4) A sample prompt using a pattern for your most common task, 5) How you'd use Research Mode in your work.
Requirements Checklist
States a specific role and 3 use-cases
Describes a Project setup with instructions
Mentions at least one MCP/tool connection
Includes a sample prompt using a pattern
References research mode or deep research
Total length is at least 200 characters