Module 1: Prompt Like a Pro
Master the art of talking to Claude to get clear, useful, and accurate responses every time.
Learn how Claude works, write structured prompts using a proven framework, create Artifacts like documents and spreadsheets, and build a fact-checking habit for reliable AI outputs.
Learning Objectives - Understand how Claude differs from a search engine and why that matters
- Write effective prompts using a structured framework to get professional-quality results
- Iterate and refine Claude outputs to match your exact needs
- Create documents, spreadsheets, and interactive tools using Claude Artifacts
- Fact-check Claude responses to ensure accuracy before sharing or acting on them
What You'll Learn - How Claude works and what makes it different from other AI tools
- The prompt engineering framework: Role, Context, Task, Format
- Iterative prompting: how to refine a vague prompt into a precise result
- Claude Artifacts: creating documents, spreadsheets, and interactive tools
- Fact-checking workflow: verifying AI outputs before you use or share them
How Claude Works
When most people first try Claude, they treat it like a faster search engine - type a question, get an answer. But Claude works in a fundamentally different way, and understanding this changes everything about how you use it.
A search engine returns links for you to visit - Claude generates a composed answer
Claude generates answers - a search engine returns links.
A search engine scans billions of existing web pages and returns links for you to visit. Claude works differently. When you send a message, Claude reads your words, understands what you are asking, and produces a composed response tailored to your situation. Where a search engine points you to sources, Claude synthesises an answer.
This is why Claude can do things no search engine can. It holds a real conversation, remembering what you said three messages ago. It reads a document you paste in and can summarise, rewrite, or analyse it. It drafts emails in your tone, explains complex topics in plain language, and helps you think through decisions step by step.
What this means for how you use Claude.
Because Claude generates responses from learned patterns rather than verified facts, it can produce inaccurate information - even confidently. This behaviour is called hallucination, and you will learn exactly how to handle it in the next section.
Claude was created by Anthropic, an AI safety company. It was trained on a large collection of text gathered up to a specific cutoff date. For current events or recent data, paste the relevant information directly into Claude so it can work with what you provide.
Watch video: How Claude Works
Key Insight: Claude produces a composed answer - it reads the sources so you do not have to. A search engine hands you links to read through yourself. Claude hands you the insight. This is what allows Claude to reason, write, and analyse in ways a search engine simply cannot.
Real-World Example: You paste a complex supplier contract into Claude and ask: "Summarise the key obligations in plain English and flag anything unusual." Claude reads every word, understands the structure, and produces a focused summary - something a search engine simply cannot do.
Q: What is the key difference between Claude and a search engine?
Claude generates a brand-new response tailored to your specific request, drawing on patterns from its training. Search engines retrieve links to existing web pages. This fundamental difference allows Claude to write, analyse, summarise, and reason in ways no search engine can.
Think about the difference between asking a search engine a question and asking a knowledgeable colleague the same question. What do you get from each? Where in your work would the colleague approach - one that thinks, writes, and reasons with you - give you the biggest advantage?
The RCTF Framework
Getting great results from Claude comes down to one thing: giving it clear instructions. Vague prompts produce generic responses - not because Claude is limited, but because it can only work with what you give it. The quality of what you get out depends almost entirely on the quality of what you put in.
The
RCTF framework is a four-part structure for writing prompts that consistently produce professional responses. RCTF stands for
Role, Context, Task, and Format. You do not need all four every time, but thinking in these terms will quickly become second nature.
Role - who you want Claude to be.
Start by telling Claude what expertise to bring. "You are an experienced HR manager" or "Act as a professional copywriter" frames everything that follows. Without a role, Claude responds as a general assistant. With a clear role, it adjusts its vocabulary, examples, and perspective to match that expertise. The same question asked to a financial analyst and to a school teacher produces very different - and both useful - results.
Context - the background Claude needs.
Claude cannot read your mind. Tell it who you are, who the audience is, what the document is for, and any relevant constraints. A prompt without context produces a generic response. A prompt with good context produces one that feels written for your specific situation.
Task - exactly what you want done.
Be precise. "Write something about onboarding" leaves too much open. "Write a 300-word welcome email for new employees joining the customer service team" is a clear task. Use action verbs: write, summarise, compare, analyse, list, explain, draft, rewrite.
Format - how you want the output.
Tell Claude how to structure the response: bullet points, a numbered list, a table, two short paragraphs, or a formal report. Without this, Claude picks a format that may not suit your workflow, and you spend time reformatting instead of using the output.
Watch video: The RCTF Framework
Key Insight: RCTF stands for Role, Context, Task, Format. Use all four in your prompts to consistently get professional-quality responses. A complete RCTF prompt takes two minutes to write and saves ten minutes of editing.
Real-World Example: Weak prompt: "Write about customer service." RCTF prompt: "You are a customer service trainer (Role). Our company handles B2B software support for SMEs (Context). Write a one-page guide on de-escalating frustrated customers (Task) as a numbered list with a short explanation for each step (Format)."
Q: What does the 'C' in the RCTF framework stand for?
C stands for Context in the RCTF framework. Context means sharing relevant background - who you are, who the audience is, what the document is for, and any constraints. Without context, Claude gives generic responses. With it, responses feel tailored to your exact situation.
Think of one task you often delegate or wish you could do faster. Write a full RCTF prompt for it. What Role would you give Claude? What Context is essential? What exactly is the Task? What Format would be most useful for how you work?
Iterative Prompting
Most new Claude users write one prompt, read the response, and if it is not perfect, they give up or start completely over. This misses what makes Claude most powerful.
Claude is built for conversation. The most effective way to use it is through a back-and-forth process called
iterative prompting - starting with a reasonable request and then refining the output through follow-up messages until you get exactly what you need.
Treat the first response as a rough draft.
Your opening prompt does not need to be perfect. Give Claude enough to produce something you can react to. Once you see the first response, you will have a much clearer picture of what to adjust. That first output is your raw material, not your final product.
Common follow-up moves.
There are reliable ways to refine any response. Change the tone: "Make this more conversational" or "This needs to sound more authoritative." Adjust the length: "Shorten this to under 100 words" or "Expand the second point with more detail." Redirect focus: "The third bullet is not relevant - replace it with a point about implementation costs." Ask for a rethink: "This is too generic. Write it specifically for a retail manager audience."
One underused technique is asking Claude: "What assumptions did you make when writing this?" Claude will often surface constraints it followed silently, which helps you write a sharper next instruction.
Context carries across the conversation.
Claude remembers everything said earlier in the same conversation. If you told it in your first message that you are writing for a senior leadership audience, it carries that forward automatically. You do not need to repeat background each time. This is what makes iterative prompting so efficient - you build progressively on what came before, shaping the final output without starting from scratch.
Key Insight: Treat Claude's first response as a rough draft, not a final product. The real power comes from follow-up messages - adjusting the tone, length, focus, and structure until you have exactly what you need.
Real-World Example: First prompt: "Write an introduction for a staff wellness policy." Claude produces a formal, wordy paragraph. Follow-up: "Good start. Shorten to 80 words and make the tone warmer." Follow-up 2: "Add one sentence explaining why wellness matters to business performance." Done in three exchanges.
Q: What is the most effective approach when Claude's first response is not quite right?
The most effective approach is to continue in the same conversation with a specific follow-up message. Claude remembers the full context and will refine, shorten, redirect, or expand based on your feedback. Starting a new conversation loses all the context you established and forces you to begin again unnecessarily.
Think of a time a tool or process did not give you what you needed on the first try. How did you course-correct? How might that same instinct apply when prompting Claude? What kinds of follow-up instructions would be most useful in your day-to-day work?
Claude Artifacts
Claude is not just a text generator you read in a chat window. It can create standalone, reusable pieces of work - formatted documents, data tables, code, and interactive tools - that appear in their own dedicated panel. These outputs are called
Artifacts.
When Claude creates an Artifact, it appears in a panel to the right of the conversation. This separates your finished work from the discussion that produced it, making it easy to review, copy, and use directly.
Types of Artifacts Claude can create.
Document Artifacts are formatted pieces of writing - reports, policies, email templates, guides, or summaries. They appear as clean, readable text you can copy directly into Word, Google Docs, or your email client.
Code Artifacts contain scripts or formulas, useful even if you are not a developer - Claude can write Excel formulas or automate simple tasks.
Spreadsheet Artifacts are structured data tables you can paste into your preferred tool.
Interactive Artifacts are small HTML applications that run right in your browser - calculators, quizzes, data entry forms, or visual comparisons built exactly to your specification.
How to request an Artifact.
You do not always need to say the word "Artifact." Simply asking Claude to create something document-like usually triggers one. Phrases that produce Artifacts include "write a report", "create a template", "draft a policy", "make a checklist", and "build a simple quiz." If Claude gives you plain inline text when you expected a formatted document, try: "Create this as an Artifact I can copy directly."
Practical uses for professionals.
Artifacts are especially useful if you create documents, templates, or tools regularly. You can ask Claude to draft a performance review form, build a client questionnaire, create a training handout, or produce a comparison table - all as ready-to-use Artifacts. The key shift is thinking of Claude not just as an answering machine but as a production tool that delivers finished work.
Watch video: Claude Artifacts
Key Insight: Artifacts appear in a separate panel to the right of your conversation. They are designed to be used directly - copy them into Word, paste into Excel, or run them in your browser - without extra formatting.
Real-World Example: You ask: "Create an onboarding checklist for new sales staff. Format it as a table with three columns: Task, Who is Responsible, and Target Completion Day." Claude produces a clean formatted table as an Artifact you can copy straight into your company template.
Q: Which of the following is NOT a type of Claude Artifact?
Claude Artifacts include formatted documents, code, spreadsheet data, and interactive HTML tools. Claude does not record or process live video calls. Artifacts are self-contained pieces of work created in a dedicated panel, designed to be copied, run, or pasted directly into your workflow.
What is one document or tool you create regularly at work that takes you a long time to format or structure? How could you ask Claude to produce it as an Artifact? What specific RCTF prompt would you write to request it?
Fact-Checking Your Outputs
Claude is a powerful tool, but it is not a source of truth. It sometimes states incorrect information with complete confidence - a behaviour called
hallucination. Understanding this and building a checking habit is one of the most important skills you will develop as a Claude user.
Hallucination happens because Claude generates responses by predicting which words and ideas fit together based on patterns in its training data. It does not have a separate fact-checking mechanism. When training data is incomplete or a topic is niche, Claude may construct a plausible-sounding answer that is factually wrong. The more specific the claim, the higher the risk.
What to always verify.
Not everything needs the same level of scrutiny. For general writing tasks - drafting emails, restructuring paragraphs, brainstorming - the risk is low. But certain categories require active verification:
statistics and numbers such as percentages, dates, and figures;
proper nouns such as company names, job titles, and legislation names;
legal and compliance claims such as rules, deadlines, and official policies; and
technical specifications such as processes and formulas. If Claude gives you a specific number or names a law, check it before you use it.
Practical fact-checking steps.
The simplest first step is to ask Claude: "Are you confident this is accurate, or should I verify it?" Claude will often acknowledge uncertainty when asked directly. Then take the specific claim to an authoritative source - a government website, official publication, or recognised industry body. For statistics, trace the original data rather than relying on a summary.
A useful habit is asking: "What sources would I search to verify this?" Claude can point you toward the right category of source, and with web search enabled it can surface sources directly. Either way, always trace the original source yourself before you send, publish, or make a decision based on AI-generated information.
Key Insight: Claude can produce incorrect information confidently - this is called hallucination. Always verify statistics, specific names, legal claims, and technical details from authoritative sources before acting on or sharing them.
Real-World Example: Claude tells you a regulation requires annual returns to be filed by 31 March. Before including this in a client document, you check the official regulatory website and find the actual deadline is 30 April. One quick check prevents a serious mistake.
Q: What is the term for when Claude states incorrect information with confidence?
Hallucination refers to when a language model generates information that sounds confident and plausible but is factually incorrect. It happens because the model predicts likely text from patterns rather than verifying facts. It is not unique to Claude - all language models do this, which is why checking specific claims from authoritative sources is a core AI literacy skill.
Have you ever acted on information that later turned out to be wrong - from a colleague, a document, or an online source? What was the cost? How will you build a fact-checking habit into your Claude workflow so you catch errors before they cause problems?
Module 2: Connect & Present
Link Claude to your work apps and turn research into stunning presentations.
Use Claude Connectors to link Google Drive, Gamma, Canva, and Dropbox. Research and summarise content from your connected data, then build polished presentations and documents with Gamma.
Learning Objectives - Connect Claude to external tools including Google Drive, Gamma, Canva, and Dropbox
- Use Claude to research, summarise, and extract insights from connected data
- Turn research and notes into polished presentations and documents using Gamma
- Build a repeatable workflow from research to final presentation
What You'll Learn - Claude Connectors: what they are and how to set them up
- Connecting Google Drive, Gamma, Canva, and Dropbox to Claude
- Researching and summarising content from connected sources
- Introduction to Gamma: AI-powered presentations and documents
- From research to presentation: a practical end-to-end workflow
What Are Claude Connectors
When you use Claude on its own, you bring information to it - you paste in text, upload files, or describe your situation. Claude Connectors flip this around: they give Claude authorised access to tools and platforms you already use at work, so it can pull in relevant information without you having to copy and paste everything manually.
A
Claude Connector is a secure integration between Claude and an external service. Once you authorise a connection, Claude can read files from your Google Drive, access your Dropbox documents, or interact with connected tools like Gamma and Canva - all within a single conversation. You stay in control: you decide which services to connect, what Claude can access, and when.
Why connectors matter for your workflow.
Without connectors, working with your own documents in Claude requires a manual step - you find the file, open it, copy the content, paste it into Claude, and then start your task. With a connector, you can simply tell Claude: "Summarise the Q3 report from my Google Drive" or "Pull the latest version of our onboarding guide from Dropbox." Claude finds the file, reads it, and works with it directly.
This is especially powerful when you are working with multiple documents. Instead of switching between tabs, copying content, and managing multiple paste operations, you can ask Claude to cross-reference several files from your connected storage, extract the relevant sections, and synthesise them into a single clear output.
Setting up connectors in Claude.ai.
Connectors are set up through the Claude.ai interface under your account settings. You authorise each service individually using your existing account credentials - Claude uses read access to retrieve files and content but does not store your data beyond the conversation. The authorisation process typically takes less than two minutes per service and only needs to be done once.
Current connectors supported by Claude include Google Drive, Dropbox, Gamma, and Canva, with more being added over time. Once connected, they are available in every Claude conversation without needing to be re-authorised.
Watch video: What Are Claude Connectors
Key Insight: Claude Connectors give Claude authorised read access to your existing work tools - Google Drive, Dropbox, Gamma, and Canva - so you can work with your documents without manually copying and pasting content into every conversation.
Real-World Example: Instead of opening a 40-page strategy document, copying the text, and pasting it into Claude, you simply say: "Summarise the key recommendations from last quarter's strategy review in my Google Drive." Claude finds it, reads it, and delivers the summary - in seconds.
Q: What is the main purpose of Claude Connectors?
Claude Connectors give Claude authorised read access to your existing work tools - Google Drive, Dropbox, Gamma, Canva, and others. This means Claude can pull in your files and documents directly without you manually copying and pasting content into every conversation.
Think about the files and documents you work with most often. Which ones would you want Claude to be able to access directly? How would removing the copy-paste step change how you work with Claude day-to-day?
Connecting Google Drive and Dropbox
Google Drive and Dropbox are two of the most common places professionals store their work files. Connecting them to Claude turns your existing document library into a searchable, workable knowledge base that Claude can draw on during any conversation.
Connecting Google Drive.
To connect Google Drive, go to your Claude.ai account settings and look for the Integrations or Connectors section. Click to add Google Drive, and you will be redirected to a standard Google authorisation screen. Sign in with your Google account and grant the permissions Claude requests - typically read access to files and folders. Once authorised, Claude can access files stored in your Drive across your conversations.
With Google Drive connected, you can reference files by name or description. Try asking: "Find my onboarding checklist in Google Drive and turn it into a step-by-step guide for new staff." Claude will locate the file, read the content, and produce the output in one smooth step. You can also ask Claude to work across multiple files: "Compare the proposals in the 'Client Pitches' folder and summarise the key differences."
Connecting Dropbox.
The process for Dropbox is similar. In the same Integrations section, select Dropbox and follow the OAuth authorisation flow. Once connected, Claude can search your Dropbox for files and read their contents on demand. Dropbox users often store reports, contracts, and reference documents here - all of which become instantly available to Claude without any manual file management.
Privacy and access controls.
Claude does not store your files or retain their content after a conversation ends. Access is read-only and scoped to what you authorise. You can revoke connector access at any time through your Claude.ai settings or through your Google or Dropbox account's connected apps page. If you work with sensitive documents, review your organisation's policies on third-party AI integrations before connecting.
Key Insight: Connecting Google Drive and Dropbox to Claude takes less than two minutes per service. Once connected, Claude can access your documents directly - no more copying and pasting content between tools.
Real-World Example: You have 12 client proposals stored in a Google Drive folder. You ask Claude: "Read all the proposals in my '2025 Pitches' folder and list the common themes across them." Claude reads each file and delivers a synthesised summary - a task that would have taken you hours manually.
Q: After connecting Google Drive to Claude, what type of access does Claude have to your files?
Claude receives read-only access to your Google Drive files - it can retrieve and work with your documents during a conversation but cannot edit, delete, or create files in your Drive. This access is scoped to what you authorise and can be revoked at any time through your Google account settings.
Which folder or set of documents in your Google Drive or Dropbox would you most want Claude to have access to? What tasks would become faster or easier if Claude could read those files directly?
Researching Connected Content
Connecting your storage is just the beginning. The real value comes when you use Claude to research, extract, and synthesise information across multiple files - tasks that normally require hours of manual reading and note-taking.
Claude reads from connected sources and generates research outputs in one workflow
How to research connected files effectively.
Start by being specific about what you are looking for. "Summarise everything in my Drive" gives Claude too much scope and too little direction. "Find the last three monthly sales reports in my Google Drive and extract the key performance trends" gives Claude a clear task with defined scope. The more specific your instruction, the more focused and useful the output.
You can also ask Claude to compare documents: "Compare the two versions of the client proposal in my Dropbox and highlight what changed." Or to extract specific information: "Find all the pricing references across the documents in my 'Contracts' folder." These cross-file tasks are where Claude with connectors truly shines.
Handling large documents.
Claude can process long documents efficiently. If you have a 50-page report, you do not need to tell Claude which pages to read - simply describe what you need: "From the annual report in my Drive, pull out all the risk factors mentioned in the executive summary and any financial appendices." Claude reads the entire document and extracts what you specified.
Building a research workflow with Claude Connectors.
The most effective approach is to think of Claude as a research coordinator rather than a search engine. Start by telling Claude what you are looking for and why - this context helps Claude filter more accurately. Then ask Claude to cross-reference findings across your Drive files to identify patterns, contradictions, or gaps. This workflow - context, search, cross-reference - takes minutes to set up and can replace hours of manual research.
Key Insight: The power of connected research is not just reading one file - it is synthesising information across multiple documents. Claude can compare proposals, extract themes from reports, and surface key facts from an entire folder of files.
Real-World Example: You need to prepare for a client meeting. You ask Claude: "Read the last four meeting notes from the 'Client ABC' folder in my Drive and summarise the outstanding action items and any unresolved concerns." Claude reads all four files and delivers a clean briefing in under a minute.
Q: What kind of research task is Claude with connected storage best suited for?
Claude with connected storage excels at synthesising information across multiple files - comparing documents, extracting themes from a folder of reports, or pulling specific data across several versions of a file. This cross-document research is where connected Claude provides the greatest time savings over manual methods.
Think about a recurring research task at work that involves reading through multiple documents - meeting notes, reports, or proposals. How would you describe that task to Claude if it had access to your stored files? What specific instruction would you give?
Introduction to Gamma
Gamma is an AI-powered tool that transforms written content into polished presentations, documents, and web pages. Where traditional presentation tools require you to manually design every slide, Gamma uses AI to generate professional layouts, suggested visuals, and clean formatting from the content you provide - in minutes rather than hours.
What Gamma creates.
Gamma can produce three main types of output.
Presentations are slide-deck style formats ideal for meetings, pitches, and training sessions.
Documents are long-form formats with sections, headings, and rich formatting - useful for reports, proposals, and guides.
Web pages are published online pages you can share via a link, suitable for landing pages, portfolios, or course materials. All three can be generated from a prompt, an outline, or a piece of text you paste in.
How Gamma integrates with Claude.
Gamma is one of Claude's official connectors, which means you can move content from Claude directly into a Gamma creation. A common workflow is to use Claude to research and draft the content - pulling from your connected files, structuring the key points, and writing the narrative - and then pass that content to Gamma to turn it into a visually designed presentation.
You can also connect Gamma from within Claude and ask Claude to generate a presentation outline that Gamma can then build from. This research-to-design pipeline removes two of the most time-consuming parts of presentation creation: finding the information and making it look professional.
Customising Gamma output.
Gamma's AI suggestions are a starting point, not a final product. You can change themes, adjust layouts, swap or remove slides, edit text, and add your own images directly within the Gamma editor. The AI handles the structural and visual heavy lifting so you can focus on the message rather than the mechanics. Gamma also supports sharing directly via link, so your audience can view the presentation in a browser without needing any software installed.
Watch video: Introduction to Gamma
Key Insight: Gamma turns written content into professionally designed presentations and documents using AI. Instead of spending hours on slide layout and formatting, you describe what you need and Gamma builds it - then you refine it.
Real-World Example: You paste the key findings from a market analysis into Gamma and prompt it: "Create a 10-slide executive presentation on our market position, aimed at the leadership team." Gamma generates a fully designed deck in under two minutes. You adjust three slides, change the colour theme to match your brand, and it is ready to present.
Q: What type of output can Gamma NOT produce?
Gamma produces presentations, documents, and web pages from written content - it does not record or generate live video. It uses AI to create professional layouts and visual formatting from the content you provide, which you can then edit and customise in the Gamma editor.
Think about the last presentation you had to create. How long did it take to gather the content, structure the slides, and make it look professional? How would a Claude-to-Gamma workflow have changed that? What would you have done with the time saved?
From Research to Presentation
One of the most powerful workflows you can build with Claude is a seamless pipeline from research to finished presentation. This end-to-end process combines Claude's ability to read, synthesise, and write with Gamma's ability to design and format - removing the two biggest drains on presentation preparation time.
The research-to-presentation workflow.
The workflow has four steps. First, connect your relevant storage - Google Drive, Dropbox, or both - to Claude in your current conversation. Second, give Claude a focused research brief: what files to read, what information to extract, and what the presentation needs to cover. Third, ask Claude to produce a structured outline or a full content draft organised into sections or slides. Fourth, take that content into Gamma and let it design the visual presentation, which you then refine and customise.
This four-step process replaces a workflow that once required you to manually search for files, read each one, take notes, draft a structure, write the slides, and then spend time on design. Claude and Gamma together handle the content extraction, structuring, and formatting - you focus on decisions, messaging, and final refinement.
Writing a clear research brief for Claude.
The quality of your presentation depends heavily on the quality of your research brief. Be specific: tell Claude what the presentation is for, who the audience is, what outcome you want from it, and which files or folders it should draw from. A brief like "Create a 10-slide executive presentation for our board meeting next Friday, drawing on the Q2 financial review and the market analysis report in my Drive, covering performance, risks, and recommended actions" gives Claude everything it needs to produce directly usable content.
Refining the output.
Using iterative prompting, review Claude's draft and adjust before passing it to Gamma. Ask Claude to sharpen a section, add a specific data point, or restructure the flow. Once the content is right, Gamma handles the design - and you adjust the visual output to match your brand and style.
Watch video: From Research to Presentation
Key Insight: The Claude-to-Gamma workflow replaces hours of manual research, drafting, and design with a focused conversation. Claude extracts and structures the content; Gamma makes it look professional; you refine and present.
Real-World Example: Brief to Claude: "Read the last three quarterly reports in my Drive and write a 10-slide presentation outline for the board, covering revenue trends, key risks, and three strategic recommendations. Format as slide titles with bullet points under each." Claude delivers the outline in minutes. You paste it into Gamma, choose a theme, and your board deck is ready.
Q: What is the correct sequence in the research-to-presentation workflow?
The correct sequence is: connect your storage to Claude, give Claude a focused research brief specifying what to read and what to cover, let Claude produce structured content or an outline, then take that into Gamma for professional design. This sequence puts content quality first and visual design second - the right order for effective presentations.
Think of an upcoming presentation or report you need to create. What files would Claude need access to? What would your research brief say? How would you describe the presentation's purpose, audience, and structure to get the best output from Claude?
Module 3: Build Your Digital Brain
Create Projects, set custom instructions, and use Skills to build a personalised AI workspace that works the way you do.
Create Claude Projects to organise knowledge, instructions, and reference files. Set up custom instructions so Claude remembers your preferences and brand voice. Use Claude Skills to add ready-made capabilities to automate repetitive tasks.
Learning Objectives - Create a Claude Project to organise knowledge, instructions, and reference files
- Write custom instructions so Claude remembers your preferences and brand voice
- Upload reference documents and files into your Claude Project
- Use Claude Skills to add ready-made capabilities to automate repetitive tasks
- Build a reusable AI workspace tailored to your specific role and work needs
What You'll Learn - What Claude Projects are and why they change how you work with AI
- Creating your first Project: name, instructions, and reference files
- Writing custom instructions: tone, format, role, and preferences
- Uploading and referencing documents inside a Project
- Claude Skills: finding and enabling ready-made capabilities for repetitive tasks
- Building a reusable AI workspace for your role
What Are Claude Projects
Every time you open a new Claude conversation, you start from scratch. Claude does not remember who you are, how you like to communicate, or what work you did together last week. For occasional tasks, this is fine. But for professionals who use Claude regularly, it means repeating context in every conversation - reintroducing yourself, explaining your role, re-specifying your preferences - before you can actually get to work.
Claude Projects solve this problem. A Project is a persistent workspace inside Claude where you store your custom instructions, reference documents, and brand guidelines once. Every conversation you open inside that Project automatically inherits all of that context. Claude already knows your role, your preferred tone, your formatting rules, and the documents you work with - before you type a single word.
A Claude Project stores your context once so every conversation starts ready to work
Why Projects change how you work with AI.
Without Projects, Claude is a general-purpose assistant you have to re-brief every time. With Projects, it becomes a specialist workspace - one that already understands your industry, your writing style, your client requirements, and the specific files relevant to your work. Over time, your Project becomes richer and more tailored, making every Claude interaction faster and more precisely suited to your needs.
Projects are especially valuable for recurring work: weekly reports written in a consistent style, client communications with a specific tone, ongoing research in a defined area, or any task where the context does not change much from session to session. Projects effectively give Claude a persistent memory for your work - something the standard chat interface does not provide by default.
Watch video: What Are Claude Projects
Key Insight: Claude Projects store your custom instructions, reference files, and brand guidelines once - and every conversation inside the Project starts with that context already loaded. No more re-briefing Claude from scratch.
Real-World Example: You create a Project called "Marketing Workspace." You add instructions: tone, writing style, target audience, brand voice. You upload your brand guidelines PDF. Now every conversation in this Project automatically writes in your brand voice without any extra prompting.
Q: What is the main problem that Claude Projects solve?
Claude Projects solve the problem of having to re-explain your role, preferences, and context at the start of every new conversation. A Project stores your custom instructions and reference files once, so every conversation inside it starts with that context already loaded - no re-briefing needed.
What context do you find yourself repeating every time you use Claude? Your role, your audience, your tone preferences? Imagine never having to type those again. What would you put in your first Claude Project?
Creating Your First Project
Setting up a Claude Project takes less than ten minutes. The process is straightforward, but the decisions you make during setup - what to call it, who it is for, and what context to include - determine how useful it becomes.
Naming your Project.
Give your Project a clear, descriptive name that tells you exactly what it is for. "Content Writing - B2B Tech" is more useful than "Writing Project 1." A good name means you can glance at your Project list and know exactly which workspace to open for each type of work. If you handle multiple clients or different types of tasks, create a separate Project for each one.
Adding a Project description.
The description is optional but useful. Think of it as a brief scope statement - what this Project is for, what types of tasks it handles, and what it should not be used for. This helps you stay consistent and makes it easier to review your Projects later.
Writing your first custom instructions.
Custom instructions are the most important part of your Project setup. This is where you tell Claude everything it needs to know to help you well - your role, your audience, your tone, your formatting preferences, and any standing rules for how you want it to respond. You will cover custom instructions in depth in the next section. For now, start with a brief version: a sentence or two about who you are and how you want Claude to write.
Uploading reference files.
You can upload documents directly into a Project - PDFs, Word documents, text files - that Claude can reference in every conversation. Start with one or two essential files: your style guide, a sample of your writing, a key reference document. You do not need to add everything at once. A small set of well-chosen files is more useful than a large collection Claude has to sift through.
Once you have created your first Project, open it and start a conversation. You will immediately notice the difference - Claude responds with the context you have set, without any prompting from your side.
Key Insight: Start simple: name your Project clearly, write a short description, add two or three lines of custom instructions, and upload one or two key reference files. You can expand and refine the Project over time as you learn what works best.
Real-World Example: You create a Project called "Client Reports - Hospitality Sector." Your custom instructions say: "I am a management consultant writing quarterly reports for hotel group clients. Use formal English, structured headings, and avoid jargon. Audience: senior hotel executives." You upload a sample previous report. Claude now writes in exactly that style for every conversation.
Q: What is the most important element to set up when creating a Claude Project?
Custom instructions are the most important part of a Claude Project setup. They tell Claude everything it needs to know about your role, your audience, your preferred tone, your formatting rules, and any standing preferences. Good custom instructions transform Claude from a generic assistant into a specialist workspace tailored to your exact needs.
Think about the type of work you do most often. What is one Project you could create that would save you the most time? What would you name it? What three things would you include in the custom instructions to get Claude writing in exactly the style you need?
Writing Custom Instructions
Custom instructions are the words you give Claude once so it remembers them always - inside your Project, for every conversation. Getting them right is the single most valuable investment you can make in your Claude setup.
Think of custom instructions as a briefing document for a new team member. You would not expect them to know your team's communication style, your clients' names, or your preferences for how reports should be structured - unless you told them. Custom instructions are that briefing, delivered once and remembered permanently.
What to include in your instructions.
Cover four key areas. First,
your role and context: who you are, what you do, and the industry or environment you work in. For example: "I am a senior HR manager at a manufacturing company with 500 employees." Second,
your audience: who you are writing for and what level of expertise they have. Third,
your tone and style: formal or conversational, concise or detailed, active or passive voice. Be specific. "Professional and approachable" is better than just "professional." Fourth,
format preferences: do you prefer bullet points or prose? Section headings or flowing paragraphs? Short responses or comprehensive ones?
Adding standing rules.
You can also include standing rules - specific things Claude should always or never do. Examples: "Always begin documents with an executive summary," "Never use jargon without defining it first," "When writing emails, keep them under 150 words unless I specify otherwise," or "Always suggest at least two alternatives when making a recommendation."
Keeping instructions clear and actionable.
Avoid vague adjectives. "Write clearly" is not actionable. "Use short sentences, one idea per paragraph, and avoid passive voice" is. The more specific and concrete your instructions, the more consistently Claude will follow them. Review your instructions after a few conversations and refine anything that is not working as expected.
Key Insight: The best custom instructions are specific and actionable. Avoid vague words like "professional" or "clear" - instead write what that means: "Use formal English, short paragraphs, and active voice. Always include a summary at the top of any report."
Real-World Example: Good custom instructions for a content marketer: "I write B2B content for a HR software company. Audience: HR managers and CHROs at mid-size companies. Tone: authoritative but approachable, no jargon. Format: structured headings, bullet points for lists, no more than 800 words per article. Always end with a clear call to action."
Q: Which of the following is the most effective custom instruction for writing style?
Effective custom instructions are specific and actionable. "Use short sentences, active voice, no jargon, one idea per paragraph, and include section headings" tells Claude exactly what to do. Vague instructions like "professional and clear" leave Claude guessing what that means in practice, leading to inconsistent results.
Take five minutes to draft your own custom instructions. Cover your role, your typical audience, your preferred tone, and at least two standing rules about format. What would you always want Claude to do? What would you never want it to include?
Uploading Reference Files
Custom instructions tell Claude how to behave. Reference files tell Claude what to know. Uploading the right documents into your Project gives Claude the specific knowledge it needs to produce output that is accurate, on-brand, and directly relevant to your work.
What types of files can you upload?
Claude Projects support a range of file formats including PDF, Word documents (DOCX), plain text files (TXT), and markdown files (MD). You can upload documents that contain your company's specific information, style guidelines, product descriptions, process documents, or any reference material that Claude would otherwise not know about.
Choosing the right files to upload.
Not every document in your organisation belongs in your Project. Choose files that Claude will genuinely need to reference regularly. Good candidates include: your brand voice and style guide, a glossary of company-specific terminology, sample documents that show the format and tone you expect, key policies or procedures Claude should be aware of, and background information about your clients or industry that Claude would not know from general training.
Avoid uploading documents just because they exist. A Project with too many files forces Claude to work through more material than necessary, which can reduce the precision of its responses. Three well-chosen files beat twenty loosely relevant ones.
How Claude uses uploaded files.
Once a file is in your Project, Claude can reference it in any conversation within that Project. You do not need to re-upload it or mention it explicitly. If you ask Claude to "write a product description in our usual style," it will draw on the style guide you uploaded. If you ask it to use "the correct terminology from our glossary," it will consult that document.
You can also reference files explicitly: "Using the brand guidelines I uploaded, rewrite this paragraph in our brand voice." This gives Claude a clear instruction to draw directly on a specific document.
Key Insight: Upload files that Claude will genuinely need to consult regularly - your style guide, a glossary, sample outputs, or key reference documents. Three well-chosen files are more useful than twenty loosely relevant ones.
Real-World Example: You upload three files to your Client Communications Project: (1) your brand voice guide, (2) a glossary of product terms, (3) three sample emails that represent the tone you want. Now every email Claude drafts for you uses your exact brand voice and correct product terminology - without you specifying this each time.
Q: Which type of file is the best candidate for uploading into a Claude Project?
The best files to upload are ones Claude will genuinely need to consult regularly and that contain information it would not otherwise know - your brand voice guide, style guidelines, glossary, or sample documents. Uploading general reference material or public documents Claude already knows wastes space and reduces precision.
Look at the documents you use most often in your work. Which three would make the biggest difference if Claude could always reference them? What information is in those documents that Claude could not know from its general training?
Automating with Claude Skills
Custom instructions and reference files make Claude smarter about who you are and how you work.
Claude Skills take this a step further - they give Claude ready-made automation capabilities so you can handle repetitive tasks in a fraction of the usual time.
A Skill is a pre-built workflow you add to your Claude workspace. Each Skill is designed for a specific job: turning your notes into a Gamma presentation, drafting a structured report from a template, generating formatted emails in your brand voice, or processing a document in a standard way. Instead of writing detailed instructions from scratch every time, you activate a Skill and it runs the workflow for you.
How to find and add Skills.
Skills are available from the Claude.ai interface. Look for the Skills section in the sidebar or explore the Skills library to browse what is available. Adding a Skill is a one-time action - once enabled, it becomes available across your conversations. Some Skills may require connecting to an external service, such as Gamma for presentations or Google Drive for document access, which you authorise during setup.
Using Skills in your workflow.
Once a Skill is enabled, you activate it from within a conversation. Provide the input it needs - a document, a set of notes, a topic, or a brief description - and the Skill runs its automation. The output appears directly in your conversation. A Skill for creating presentations might take your research notes and produce a structured Gamma deck. A document Skill might reformat a long report into a clean summary in your standard template.
You do not need technical knowledge to use Skills. They are designed to be activated with a simple instruction and do the rest automatically.
Skills inside a Claude Project.
Skills become particularly powerful when combined with a Claude Project. When you add a Skill to a Project, it sits alongside your custom instructions and reference files - available in every conversation within that workspace. This means your Project not only remembers your preferences and documents, but also has your most-used automations ready to run without any setup.
The result is a fully equipped AI workspace: Claude knows who you are, writes the way you prefer, references the documents you need, and can automate the tasks you repeat most often. This combination - Projects plus Skills - is where Claude shifts from a helpful assistant into a genuine part of how you work.
Key Insight: Claude Skills are pre-built automations that handle specific repetitive tasks without you writing instructions from scratch. Add the Skills that match your most common workflows and they are ready to run whenever you need them.
Real-World Example: You regularly turn meeting notes into a formatted status report. Instead of prompting Claude from scratch each time, you enable a document Skill built for this workflow. Next time, you paste your notes, activate the Skill, and receive a correctly formatted report in under a minute - every time, in exactly the same structure.
Q: What is the main purpose of Claude Skills?
Claude Skills are pre-built automation capabilities you add to your workspace to handle specific repetitive tasks. Rather than writing instructions from scratch each time, you activate a Skill and it runs the workflow automatically. Skills are designed for particular jobs - presentations, document formatting, communications - so you get consistent, fast results without manual prompting.
Think about the tasks you repeat most often that follow roughly the same steps each time - turning notes into reports, drafting a certain type of email, formatting a document in a standard way. Which of these would you want a ready-made Skill to automate? What would you give it as input, and what would you expect it to produce?
Building Your Reusable AI Workspace
Once you have your first Project set up with custom instructions and a few reference files, you have created something genuinely valuable: a personalised AI workspace that gets more useful the more you use it.
But most professionals benefit from more than one Project. The goal is to build a small portfolio of purpose-built workspaces - each one tuned for a specific type of work - so that whatever you are doing, you have a Claude setup already primed for it.
Designing your Project portfolio.
Think about the different types of work you do and the different contexts you operate in. A marketing manager might have Projects for content writing, client proposals, social media copy, and internal communications. A consultant might have separate Projects for each major client. A trainer might have Projects for each course they run.
The key is that each Project has a distinct purpose and audience. When the purpose and audience change, the custom instructions and reference files should change too. Mixing too many contexts into one Project dilutes its usefulness.
Maintaining and improving your Projects.
A Project is not set-and-forget. Review your custom instructions every few months and update them if your role, clients, or preferences have changed. Add new reference files as they become relevant. Remove files that are no longer current. The more accurately your Project reflects your current work, the more effectively Claude performs inside it.
Sharing insights across Projects.
If you find a custom instruction, reference file, or Skill that works particularly well in one Project, consider adapting it for others. Over time, you build an evolving library of AI workspace configurations - each one refined through real use - that represents your personal way of working with Claude. As your work evolves, your Workspace evolves with it - update the instructions, add new files, enable new Skills, and adjust your preferences at any point to keep Claude aligned with where you are.
Key Insight: Build a portfolio of purpose-built Projects, each with tailored instructions, the right reference files, and relevant Skills. This trio makes each Project a fully equipped specialist workspace - ready to go without any briefing.
Real-World Example: A HR consultant builds three Projects: "Job Descriptions" (formal HR language, structured format, skills-focused), "Employee Communications" (warm, inclusive, jargon-free), and "Client Proposals" (business formal, ROI-focused, executive audience). Each Project has its own custom instructions, reference files, and relevant Skills - making Claude an instant specialist in each context.
Q: Why is it better to build multiple focused Projects rather than one general workspace?
Multiple focused Projects work better than one general workspace because each Project is tailored to a specific context - its own custom instructions, reference files, Skills, and purpose. A Project for client proposals behaves differently from one for internal reports, because the tone, audience, and goals are different. Tailoring each Project makes Claude more precisely useful for each type of work.
How many different types of work do you do regularly that might benefit from their own Claude Project? Sketch out two or three Project names. For each one: what custom instructions would you write, which reference files would you include, and which Skills would save you the most time?
Module 4: Use Claude Everywhere
Bring Claude into Microsoft 365, your browser, and your daily work routine.
Install Claude for Microsoft 365 to work smarter inside Word, Excel, and PowerPoint. Add the Claude in Chrome extension to summarise web pages as you browse, and set up basic automation for everyday tasks.
Learning Objectives - Install and use Claude for Microsoft 365 inside Word, Excel, and PowerPoint
- Install the Claude in Chrome extension and summarise web content as you browse
- Use Claude in Chrome to research and extract information across multiple tabs
- Set up a simple WhatsApp automation using Claude in Chrome
What You'll Learn - Claude for Microsoft 365: installation and setup
- Using Claude inside Word, Excel, and PowerPoint
- Claude in Chrome: installation and core features
- Summarising and researching web pages with Claude in Chrome
- Basic WhatsApp automation using Claude in Chrome
Claude for Microsoft 365
Most professionals spend a significant part of their working day inside Microsoft Office applications - writing in Word, analysing data in Excel, building slides in PowerPoint. Claude for Microsoft 365 brings AI assistance directly into these tools, so you do not need to copy content out of your documents and paste it into a separate Claude window.
Claude for Microsoft 365 brings AI assistance into Word, Excel, and PowerPoint
What Claude for Microsoft 365 provides.
Claude for Microsoft 365 appears as an add-in within your Office applications. You activate it from the ribbon or toolbar - depending on the app - and a Claude panel opens alongside your document. From there, you can ask Claude to help with whatever you are working on without leaving the application.
The integration is contextual: Claude can see and read the document you are working on (with your permission), which means you can ask it to review what you have written, suggest improvements, or continue writing in the same style. You do not need to copy and paste - the content is already available.
Installing Claude for Microsoft 365.
The Claude for Microsoft 365 add-in is available through the Microsoft AppSource store - the same place you find other Office add-ins. Open any Microsoft 365 app, go to Insert, then Add-ins, and search for Claude. Click to install and authorise it with your Claude account. Once installed, it appears in your ribbon across all your Microsoft 365 applications on that device.
Anthropic periodically updates the integration, so the exact steps may vary slightly. If you encounter any issues, check the Anthropic website for the latest installation guide.
Watch video: Claude for Microsoft 365
Key Insight: Claude for Microsoft 365 puts AI assistance directly inside Word, Excel, and PowerPoint - no more switching between apps or copy-pasting content. You can ask Claude to help while the document is open in front of you.
Real-World Example: You are writing a staff policy in Word. You highlight a dense, formal paragraph and ask Claude in the sidebar: "Rewrite this in plain English for non-HR staff." Claude rewrites it within Word without you opening a single other window.
Q: Where do you install the Claude for Microsoft 365 add-in?
The Claude for Microsoft 365 add-in is installed through the Microsoft AppSource store. Open any Microsoft 365 app, go to Insert, then Add-ins, and search for Claude. Once installed and authorised, it appears in your ribbon across all your Microsoft 365 applications on that device.
Think about the Microsoft 365 tasks that currently take you the most time - drafting reports, cleaning up presentations, or making sense of spreadsheet data. Which one would benefit most from having Claude directly inside the application? What would you ask it to do first?
Claude in Word, Excel, and PowerPoint
Each Microsoft 365 application has different strengths, and Claude integrates with each one in ways that match the work you do inside it. Understanding what Claude can do in each app helps you get the most out of the integration.
Claude in Microsoft Word.
Word is where most writing happens - policies, reports, proposals, emails, and documentation. With Claude available in the sidebar, you can ask it to review a document for clarity, rewrite sections in a different tone, extend a draft that has stalled, or produce a summary of a long document without leaving the file. You can also ask Claude to check your writing for a specific audience: "Is this clear enough for a non-technical reader?" or "Does this meet the tone we use for external communications?"
Claude in Microsoft Excel.
Excel users without a technical background often struggle with formulas, functions, and data interpretation. Claude in Excel can explain what a formula does in plain English, suggest the right formula for a task you describe, help you understand patterns in your data, or write a commentary explaining what a set of figures means. You can ask: "What does this VLOOKUP formula do?" or "I have monthly sales figures - what is the best way to show the trend over the year?"
Claude in Microsoft PowerPoint.
PowerPoint presentations often suffer from too much text, unclear structure, and weak narratives. Claude can help you improve each aspect without rebuilding the deck from scratch. Ask it to tighten the bullet points on a slide, write speaker notes for each page, suggest a clearer order for your argument, or rewrite a title to be more impactful. You can also ask Claude to review the deck as a whole: "Does the flow of this presentation make sense for a board audience?" and get structural suggestions before you present.
Key Insight: In Word, Claude helps you write and rewrite. In Excel, it helps you understand and explain data. In PowerPoint, it helps you tighten the message and improve structure. Each application gives you a different reason to use Claude in context.
Real-World Example: You have 12 slides in PowerPoint but the narrative does not flow. You ask Claude: "Review these slide titles in order and tell me where the argument loses coherence." Claude identifies two slides that break the logical flow and suggests a reorder - in two minutes, without redesigning anything.
Q: Which of the following is the best use of Claude inside Microsoft Excel?
Claude in Excel is most valuable for helping non-technical users understand and work with formulas, interpret data patterns, and get suggestions for functions that match a task they describe. Visual redesign, database imports, and transcription are not core Claude-in-Excel use cases.
Which of the three Microsoft 365 apps - Word, Excel, or PowerPoint - would help you most if you had Claude available inside it right now? What specific task would you try first, and what would you ask Claude to do?
Claude in Chrome
A significant amount of professional knowledge work happens in a web browser. You research topics, read articles, review websites, fill in online forms, and manage cloud tools - all inside Chrome. The Claude in Chrome extension brings Claude directly into your browsing experience, so you can ask for help with whatever is on your screen without switching to a separate tab.
What Claude in Chrome does.
Claude in Chrome is a browser extension that adds a Claude panel to your Chrome browser. Once installed, you can open it at any time while browsing - typically by clicking the Claude icon in your browser toolbar or using a keyboard shortcut. Claude can see the content of the web page you are currently viewing (with your permission), which means you can ask it questions about that page without manually copying text.
Key features include one-click page summarisation, where Claude reads the current page and distils the key points into a concise summary. You can also ask follow-up questions about the page content, extract specific information - such as "list all the pricing tiers mentioned on this page" - or ask Claude to compare what you are reading with your own knowledge or notes.
Installing Claude in Chrome.
Install the Claude in Chrome extension from the Chrome Web Store. Search for "Claude for Chrome" or "Claude AI" and click Add to Chrome. Once installed, sign in with your Claude account to activate it. The extension appears as an icon in your Chrome toolbar. You may need to pin it to keep it visible by right-clicking the icon and selecting Pin.
The extension works on any web page in Chrome. Some pages with restricted content or protected viewing may limit what Claude can read, but for most everyday browsing - news articles, LinkedIn posts, websites, online documents - it works seamlessly. For most professionals, the Chrome extension becomes the default starting point for any web research task - open the sidebar first, then navigate to the page you need.
Watch video: Claude in Chrome
Key Insight: Claude in Chrome brings AI assistance directly into your browser. With one click, Claude can summarise the web page you are reading, extract key facts, or answer questions about the content - without you copying a single word.
Real-World Example: You are reading a 3,000-word industry report online. You open Claude in Chrome and ask: "Summarise the three most important findings from this report and any specific data points mentioned." Claude reads the page and delivers a focused summary in seconds - no copying, no pasting, no new tab required.
Q: What does Claude in Chrome allow you to do when browsing a web page?
Claude in Chrome lets you ask questions about or get summaries of the web page you are currently viewing - without opening a new tab or copying any text. Claude reads the current page content (with your permission) and provides analysis, summaries, or specific extractions on demand.
Think about the web pages you spend the most time reading for work - reports, news, product pages, competitor sites. How would the ability to instantly summarise or extract key data from any of those pages change how you gather information?
Research Across Multiple Tabs
One of the most powerful - and underused - features of Claude in Chrome is the ability to research across multiple open tabs. Instead of reading five articles, taking separate notes, and manually synthesising the information, you can gather your sources and ask Claude to do the analytical work for you.
How multi-tab research works.
With Claude in Chrome open, you can switch between tabs and ask Claude to read each one in sequence, building up a picture of the topic across multiple sources. Some versions of the extension allow Claude to access all currently open tabs at once. In others, you navigate to each tab and paste a summary request before moving to the next. The end result is the same: Claude synthesises across multiple sources to give you a cross-referenced output.
A practical approach is to open three to five relevant articles, competitor pages, or reports in separate tabs. Then ask Claude: "I have several tabs open about this topic. Summarise the common themes and any significant differences between the sources." Claude reads the available content and produces a synthesised view that would have taken you 30 to 60 minutes to compile manually.
Practical applications at work.
Multi-tab research is especially useful for competitive research - reading several competitor websites and extracting positioning differences. It is also valuable for due diligence, where you need to cross-reference information from multiple sources before making a recommendation. Professionals who need to stay current with industry news use it to process several articles at once and extract what is relevant to their specific role.
Ask Claude specific questions rather than open-ended ones to get the most useful output. "What do these sources say about pricing trends?" produces a more actionable summary than "summarise all of this." The more focused your question, the more directly useful Claude's synthesis will be.
Key Insight: Multi-tab research turns an hour of manual reading and note-taking into a few minutes of focused conversation with Claude. Open your sources, ask targeted questions, and let Claude synthesise the information across all of them.
Real-World Example: You need to prepare a competitive brief. You open the websites of five competitors in separate tabs. You ask Claude: "Based on the pages I have open, summarise how each competitor positions themselves and what they claim as their main differentiator." Claude reads each page and delivers a side-by-side comparison you can use immediately.
Q: What is the main advantage of using Claude in Chrome for multi-tab research?
The main advantage is that Claude can synthesise information across multiple web sources, doing in minutes what would take 30 to 60 minutes of manual reading and note-taking. You ask Claude targeted questions across your open tabs and it delivers a cross-referenced, synthesised output you can act on directly.
What is one research task you do regularly that involves reading multiple websites or articles and pulling out the key information? How would you set up your tabs and what specific question would you ask Claude to synthesise across them?
WhatsApp Automation with Claude
Claude in Chrome can help with more than just web research. Because Chrome is where many professionals manage their WhatsApp Web conversations, Claude can assist with one of the most time-consuming parts of daily communication: drafting clear, appropriate replies at speed.
Using Claude with WhatsApp Web.
WhatsApp Web runs in a browser tab, which means Claude in Chrome can see the conversation you are looking at (with your permission). This creates a simple workflow: you open a WhatsApp conversation, review the messages you need to respond to, and ask Claude to help draft a reply. Claude reads the context of the visible conversation and produces a response that matches the tone and subject matter.
For example: you have a client who has sent a long, detailed message with several questions. Instead of reading each question separately and composing answers one by one, you ask Claude: "Help me draft a professional reply to this WhatsApp message. Address each question clearly and keep the tone warm but formal." Claude reads the conversation and produces a draft that you can then review, adjust, and send.
Automating response patterns.
For recurring types of messages - common customer questions, standard status updates, or routine follow-ups - you can build a simple workflow. Keep Claude open in a split screen alongside WhatsApp Web. When a familiar type of message arrives, paste the key details into Claude and ask for a response that follows your usual pattern. Claude produces a draft in seconds that you edit and send.
This is not full automation in the technical sense - you are still reviewing and sending each message manually. But it significantly speeds up high-volume communication by removing the time you spend staring at a blank reply box deciding how to phrase a response. Over a working day, this kind of micro-automation adds up to a meaningful saving of time and mental energy.
Key Insight: Claude in Chrome can see your WhatsApp Web conversations and help you draft replies on the spot. For high-volume communication, this turns minutes of careful composing into seconds of reviewing a Claude draft and clicking send.
Real-World Example: A client sends a WhatsApp message with four questions about a project. You open Claude in Chrome, point it at the conversation, and ask: "Draft a professional reply in a warm but businesslike tone. Answer all four questions concisely." You review the draft, make one small adjustment, and send - in under 90 seconds.
Q: How does Claude in Chrome help with WhatsApp communication?
Claude in Chrome reads your visible WhatsApp Web conversation and helps you draft a reply that addresses the context and questions in the message. You still review, adjust, and send the reply manually - Claude speeds up the drafting step, not the sending step. This is particularly useful for complex or high-volume messages.
Think about the types of messages or emails you find hardest or most time-consuming to reply to. Which ones would benefit most from having Claude help you draft a response? What instruction would you give Claude to match your usual communication style?
Module 5: Meet Your AI Coworker
Take Claude off the browser and onto your desktop, and take your first steps with AI agents that work independently on your behalf.
Install Claude Desktop to work with your local files, discover Claude Cowork for multi-step task delegation, review the complete Claude AI ecosystem, and prepare for the Claude Creator Program.
Learning Objectives - Download and set up Claude Desktop for Windows or Mac
- Use Claude Desktop for everyday tasks involving local files
- Understand what Claude Cowork is and how it handles multi-step tasks
- Review all five layers of the Claude AI ecosystem
- Understand the pathway to the Claude Creator Program and Demo Day
What You'll Learn - Claude Desktop: the standalone app and what makes it different
- MCP: how Claude Desktop connects to local files and tools
- Everyday task delegation with Claude Desktop
- Introduction to Claude Cowork: AI agents and multi-step tasks
- The complete Claude AI ecosystem: five layers reviewed
- The Claude Creator Program: what comes next
Claude Desktop
So far in this programme, you have used Claude through a web browser. Claude Desktop is a standalone application for Windows and Mac that runs outside your browser. This small change opens up a much larger set of possibilities.
Claude Desktop extends the browser version with local file and system access
The core advantage: local access.
When you use Claude in the browser, every piece of information you give Claude must be manually pasted or uploaded into the conversation. Claude Desktop changes this. Once installed, Claude Desktop can read files stored on your computer - documents, spreadsheets, notes - without you having to copy and paste them first. This saves significant time when you regularly work with local files.
Claude Desktop also supports a technology called MCP (Model Context Protocol), which allows Claude to connect to external tools and services installed on your machine. This is the foundation for more advanced automation workflows that you will explore in the Claude Creator Program.
Who should use Claude Desktop.
If most of your work happens in the browser and you do not store files locally, the browser version of Claude is perfectly sufficient. Claude Desktop is most useful for professionals who regularly work with files on their computer - drafting reports, reviewing documents, managing local project folders. It is also the gateway to more advanced uses involving MCP tools, which open up automation possibilities well beyond what the browser alone can do.
Claude Desktop is available for Windows and macOS and is free to download for all Claude users.
Once installed, Claude Desktop runs as a standalone application outside your browser. You can switch between it and your other apps just like any other program on your computer, making it easy to incorporate into an existing work routine.
Watch video: Claude Desktop
Key Insight: Claude Desktop can read files on your local computer without you needing to copy and paste them. For anyone who works with documents saved on their machine, this alone makes it worth installing.
Real-World Example: You have a folder of client meeting notes saved on your desktop. With Claude Desktop, you can ask Claude to read through all of them and produce a summary of the key action items - without opening each file manually.
Q: What is the main advantage of Claude Desktop over the browser version of Claude?
Claude Desktop's key advantage is its ability to read files stored locally on your computer. The browser version requires you to manually paste or upload content, while Claude Desktop can access local files directly - saving significant time for users who regularly work with documents on their machine.
Think about the files you work with most often - reports, notes, spreadsheets. How might being able to give Claude direct access to those files change how you work?
Everyday Tasks with Claude Desktop
Once Claude Desktop is installed, you can start using it with your real work files immediately. The workflow is straightforward: open Claude Desktop, start a conversation, and when you need Claude to work with a local file, drag the file into the chat or use the attachment button to select it from your computer. Claude reads the full content of that file and works with it directly.
How professionals use Claude Desktop day-to-day.
The most immediate benefit is time saved on document-heavy work. If you receive a lengthy report and want a summary, you can hand the file directly to Claude Desktop rather than copying and pasting sections. If you have a folder of client feedback documents and want to identify recurring themes, Claude Desktop can read through them and synthesise the patterns.
Three practical scenarios where Claude Desktop makes the biggest difference:
Morning briefing from local notes. Many professionals keep running notes on their computer. With Claude Desktop, you can ask Claude to read your notes from the past week and prepare a concise briefing of your open tasks and upcoming priorities.
Summarise your document library. If you have a collection of reports, contracts, or reference materials saved locally, Claude Desktop can read through multiple documents in a session and extract the key points from each. This is considerably faster than manually reviewing each file.
Draft from local source material. When writing a proposal or presentation, you can point Claude Desktop to existing documents on your computer as source material. Claude reads them and produces a draft that incorporates your own content, rather than starting from a blank page.
Claude Desktop and the browser version side by side.
You do not have to choose between Claude Desktop and the browser version. Many professionals use both: Claude in the browser for quick questions and prompts, and Claude Desktop when working on something that requires pulling in local files. The Projects you have built and the Artifacts you have saved in the browser version remain accessible in the browser, so both tools have their place in a well-rounded AI workflow.
Key Insight: You do not need to change your entire workflow to benefit from Claude Desktop. Start with the task where you spend the most time copying and pasting content into Claude, and replace that manual step with a local file attachment.
Real-World Example: A consultant has 12 client interview transcripts saved locally. Instead of pasting each one into the browser, she drags all 12 into Claude Desktop and asks: "Read through these transcripts and identify the five most common pain points clients mentioned." Claude reads them all and returns a ranked list with supporting quotes.
Q: Which of these is a key advantage of using Claude Desktop over the browser version?
The key advantage of Claude Desktop is the ability to read local files directly. In the browser version, you must manually copy and paste content from your documents into the conversation. Claude Desktop removes this step by connecting to files on your computer, which saves significant time for anyone who regularly works with locally-saved documents.
Which files or documents on your computer do you spend the most time copying into Claude? How would direct file access through Claude Desktop change that workflow?
Introduction to Claude Cowork
Claude Cowork is a feature that changes how you can delegate tasks to Claude. In a standard Claude conversation, Claude responds to each of your messages one at a time. You ask a question, Claude answers, you ask a follow-up, Claude responds again. The back-and-forth is immediate but requires your continuous involvement.
Claude Cowork introduces a different model: delegate and review. Instead of a real-time conversation, you give Claude a multi-step task with a clear goal. Claude then works through the steps, and you return when it has finished to review the result. You are not absent from the process entirely. Claude can check in at key decision points, but the routine steps happen without you needing to respond to each one.
What makes Cowork different from a standard chat.
The best way to understand Claude Cowork is to think of it as assigning work to a capable assistant rather than having a conversation. You might say: "Research the top five competitors in our market. For each one, summarise their pricing, their main offering, and one thing they do better than us. Compile the results into a table." In a standard chat, Claude would respond once with whatever it could manage from memory. With Cowork, Claude can carry out the research steps and return a properly compiled result.
What Claude Cowork is good for.
Cowork is most valuable for tasks that have multiple sequential steps, where you do not need to be involved in each step, but you do want to review and approve the final output. Research and summarisation tasks are a natural fit. So are tasks that require gathering information from multiple sources and organising it in a structured way.
An early-access feature.
Claude Cowork is currently in early access. Not all users have it available yet, and the interface and capabilities continue to evolve. The core principle is stable: instead of continuous conversation, you delegate a task and return to review the outcome. This is the direction Claude is moving as it becomes capable of longer, more autonomous work.
Watch video: Introduction to Claude Cowork
Key Insight: Claude Cowork is not a replacement for the Claude chat you already know. It is a complementary mode for longer, multi-step tasks where you want to set a goal and return to review results, rather than staying engaged throughout.
Real-World Example: You need to prepare a briefing on five fintech startups for a client meeting. You give Claude Cowork the task: research each company, find their latest funding round, key product, and main differentiator. Claude works through all five and returns a structured briefing document for you to review and edit.
Q: What best describes how Claude Cowork differs from a standard Claude conversation?
Claude Cowork uses a delegate-and-review model. You assign a multi-step task, Claude works through it, and you return to review the output. This is fundamentally different from standard chat, where you are involved in each turn of the conversation. Cowork is not a team feature and does not require a higher-tier subscription - it is a different mode of working with Claude on longer tasks.
Think of a recurring research or information-gathering task in your work. How might delegating that task to Claude Cowork, rather than doing it step-by-step in a conversation, change the result or the time you spend on it?
Your Complete AI Ecosystem
By reaching this final module, you have built up a complete picture of how Claude works as a professional tool. It is worth pausing to see everything as a connected whole, because the layers you have learned each build on the ones before them.
The five layers of the Claude ecosystem you have built across this programme
Layer 1: Prompting.
Every interaction with Claude starts with a prompt. In Module 1, you learned how to write prompts that get consistently useful results. This is the foundation. Without clear prompting, no other feature in the Claude ecosystem produces good work.
Layer 2: Connectors.
In Module 2, you connected Claude to the tools you already use: Google Drive, Gamma for presentations, Canva for design, Dropbox for file storage. Connectors remove the manual step of copying content in and out of Claude, and enable you to produce real outputs like slides and documents directly.
Layer 3: Projects and Skills.
In Module 3, you built a Claude Project and added Claude Skills. Projects give Claude persistent memory, consistent instructions, and access to the documents that matter for a specific topic or client - so Claude always knows your context. Skills add ready-made automation capabilities for repetitive tasks, so common workflows run in one step rather than being prompted from scratch each time.
Layer 4: Everywhere.
In Module 4, you brought Claude into Microsoft 365 and Chrome. This means Claude is available inside Word, Excel, PowerPoint, and your browser. You do not have to leave the application you are working in to get AI assistance.
Layer 5: Desktop and Cowork.
In this final module, you have gone further: Claude Desktop gives you direct access to files on your computer, and Claude Cowork introduces the idea of delegating multi-step tasks and reviewing results. These represent the next frontier of how professionals work with AI.
Key Insight: Most professionals use Claude only at the prompting level. Completing all five layers of this programme puts you significantly ahead of the majority of AI users in any professional environment.
Real-World Example: A project manager starts with prompting (Layer 1) to draft meeting summaries. She adds Connectors (Layer 2) to pull from Google Drive. She builds a Project and adds relevant Skills (Layer 3) for her main client. She uses Claude in Word (Layer 4) to draft reports. Finally, she uses Claude Desktop (Layer 5) to review local project files. Each layer built on the last.
Q: What does Layer 3 (Projects) add to the Claude AI ecosystem?
Claude Projects and Skills (Layer 3) together create a personalised AI workspace. Projects provide persistent memory - Claude remembers your instructions, files, and preferred style. Skills add ready-made automations for repetitive tasks. Layer 2 covers Connectors (Google Drive, Gamma, Canva), Layer 4 brings Claude into Microsoft 365 and Chrome, and Layer 5 covers Desktop and Cowork.
Looking at the five layers of the Claude ecosystem, which layer do you feel most confident using? Which layer represents the biggest jump forward in how you currently work with AI?
Ready for the Creator Program
You have now completed all five modules of the Claude AI Program. You have learned to write effective prompts, connect Claude to the tools you use every day, build a personalised AI workspace with Projects and Skills, bring Claude into Microsoft 365 and your browser, and take your first steps with Claude Desktop and Claude Cowork. That is a substantial body of knowledge built across a short programme.
What the Claude Creator Program covers.
The Claude Creator Program is the Level 2 continuation, designed for people who want to go beyond using Claude and start designing with it. Where the Claude AI Program focused on using the Claude tools that already exist, the Creator Program focuses on building: designing AI workflows for specific business problems, creating custom instructions and prompt systems, and building simple tools for yourself or your team.
The Creator Program is not a technical course. You do not need to know how to code. Instead, it is about thinking strategically: understanding what problem you are solving, designing the AI workflow that addresses it, and building a solution that others can use.
The capstone project and Demo Day.
Every participant in the Creator Program completes a capstone project: a real AI workflow or tool built around an actual challenge from their work. At the end of the programme, participants present their project on Demo Day, a live session where each person shows what they built and how it works.
Demo Day is not a test. It is a showcase. You will see what others have built and share what you have made. Participants consistently say Demo Day is the most valuable session in the entire programme, because seeing different approaches to the same challenge generates ideas you would not have found on your own.
Certificate of Achievement.
Participants who complete the Creator Program and present on Demo Day receive a Certificate of Achievement from AICoach.my. This recognises both the learning and the practical work of building and presenting a real AI solution. If you are ready to go further with AI, the Creator Program is the natural next step.
Key Insight: The Claude Creator Program is designed for professionals who want to build AI solutions, not just use them. No coding required. The focus is on designing workflows, solving real problems, and creating tools others can use.
Real-World Example: A trainer who completes the Creator Program builds a custom AI intake process for new clients: a structured prompt sequence that gathers their goals, learning style, and context. This is not coded software. It is a designed workflow built entirely in Claude. She presents it on Demo Day and earns her certificate.
Q: What do participants present on Demo Day at the end of the Claude Creator Program?
Demo Day is a live showcase where each Creator Program participant presents the capstone project they built: a real AI workflow or tool designed to address an actual challenge from their work. It is not a test or exam. The focus is on showing what you built and how it works. Participants consistently rate Demo Day as the most valuable session in the programme.
If you were to build one AI workflow for your work right now, using everything you have learned in this programme, what problem would you tackle first?
