Personal Productivity System with Claude

Build a personalized Claude workspace with projects, memory, and reusable tools.

Introduction

30 Second Summary

Think about how much time you spend each week switching between apps, re-explaining your situation to different tools, and starting from scratch every time you need help with a decision or plan. What if one assistant already knew your goals, your style, and your preferences?

In this project, you will build a complete personal productivity system inside Claude that remembers who you are, organizes your work into dedicated projects, and produces interactive tools you can reuse every day.

What You'll Build

A personalized Claude workspace with custom instructions, uploaded reference files, and reusable interactive artifacts that solve real tasks in your life.

By the end of this project, you'll have:

A fully configured Claude account with memory, code execution, and web search enabled.
A Personal Productivity Project with custom instructions and uploaded reference material that shapes every response.
A collection of interactive artifacts including a weekly planner, a decision-making framework, and a polished downloadable document.
Secret Mission: Build an AI-powered coaching artifact that asks you questions and delivers personalized advice.

Are there any prerequisites?

None. You do not need any coding experience or prior AI knowledge. This project uses Claude's free web interface at claude.ai, and everything you need is set up in Step 1.

Set Up Your Claude Account

Before you can build your productivity system, you need a workspace to build it in. Claude is a free AI assistant that lives in your browser, and it comes with powerful features like memory, file creation, and web search built right in.

In this step, you'll create your free account, flip on the key capabilities, and send your very first message to confirm everything is working.

In this step, get ready to:

Create your free Claude account and explore the interface.
Enable Memory, Code execution and file creation, and Web search.
Send your first message and confirm Claude responds.

Create your free Claude account

Claude runs entirely in a web browser. You'll sign up at claude.ai and land in the main chat interface within a couple of minutes.

Open your web browser and navigate to:
https://claude.ai
Click Sign Up on the landing page.
Choose a sign-up method (Google account or email address) and follow the prompts to create your free account.
Once signed in, you'll land on the main chat screen.

Take a moment to look around the interface. Notice the sidebar on the left (where your conversations will live), the large chat area in the center (where you'll type messages), and your initials in the lower left corner (where you'll access settings).

What is Claude?

Claude is an AI assistant made by Anthropic. The free plan gives you access to chat on web and mobile, web search, memory across conversations, code execution, file creation, and extended thinking. No credit card is required.

Can't create an account?

Make sure you're visiting claude.ai (not the API console). If you're asked to verify a phone number, follow the on-screen instructions. Claude accounts are available worldwide, but some regions may require phone verification.

help me create my Claude account

Enable key capabilities

Claude's most powerful free features are toggled off by default. You'll turn on three capabilities that the rest of this project depends on: Memory, Code execution and file creation, and Web search.

Click your initials in the lower left corner of the screen.
Select Settings from the menu that appears.
Click Capabilities in the left sidebar of the Settings page.
Find the Memory toggle and switch it on.
Find the Code execution and file creation toggle and switch it on.
Find the Web search toggle and switch it on.

Why turn these on now?

Memory lets Claude remember context across conversations, so it can learn your preferences over time. Code execution and file creation lets Claude generate downloadable documents like PDFs, Word files, and spreadsheets. Web search lets Claude pull up-to-date information from the internet when answering your questions.

All three are included free on your plan. Turning them on now means every feature in this project will work without interruption.

Don't see a toggle?

Make sure you clicked your initials (lower left corner) then selected Settings, and that you're on the Capabilities tab. If a toggle appears grayed out, your organization admin may have disabled it. Individual free-plan users should see all three toggles available.

help me find the Capabilities settings in Claude

Send your first message

With your capabilities enabled, it's time to confirm everything works by having your first conversation with Claude.

Click the Claude logo or New chat to start a fresh conversation.
Type the following message into the chat box and press Enter to send it:

Hi Claude! What model are you, and what capabilities do you have enabled right now?

Wait for Claude to respond. It should identify itself and mention its available capabilities.

You should see Claude respond within a few seconds. Look for a mention of Sonnet 4.6 in the response or in the model indicator near the chat input. This is the default model on the free tier.

What is Sonnet 4.6?

Claude Sonnet 4.6 is the default model for free and Pro plans on claude.ai. It handles everyday tasks like writing, analysis, and conversation well. You don't need to change it for this project.

Not seeing a response?

If Claude takes more than 30 seconds, check your internet connection and try refreshing the page. The free plan has limited usage capacity, so if you see a rate limit message, wait a few minutes and try again.

help me troubleshoot why Claude isn't responding

Your Claude account is ready to go. Next up, you'll learn how to write prompts that get dramatically better results from Claude.

Master Prompt Engineering Basics

Your Claude account is set up and ready to go. Now it's time to learn the skill that separates casual users from power users: how to write prompts that get exactly what you need.

The quality of Claude's responses depends entirely on how you ask your questions. In this step, you'll see the difference firsthand by testing the same request written two different ways.

In this step, get ready to:

Test the difference between vague and specific prompts.
Practice three core prompting techniques that improve every response.
Write a structured prompt using XML tags for complex requests.

Test vague versus specific prompts

Anthropic's own guidance says to treat Claude like a newly-hired contractor with no context. The more specific your instructions, the better the result.

You'll test this by sending the same request two ways and comparing what comes back.

Start a new conversation in Claude by clicking the compose button in the sidebar.
Type this vague prompt into the chat and press Enter:

help me plan my week

Read through Claude's response. It will be generic and broad because Claude has no idea who you are, what your week looks like, or what format you want.

Now paste this specific prompt into the same conversation and press Enter:

I'm a marketing manager who works 9-5, Monday to Friday. I have 3 major deadlines this week: a campaign launch on Wednesday, a client presentation on Thursday, and a quarterly report due Friday. I also need to fit in 30 minutes of exercise each day.

Please create a day-by-day plan for my week in a table format with columns for Time, Task, and Priority Level (high/medium/low).

What changed between these two prompts?

The second prompt gives Claude three things the first one lacked: context about who you are, constraints about your schedule, and a specific output format (a table with named columns).

Claude went from guessing what you might need to producing a structured, actionable plan. The prompt did the heavy lifting.

Compare the two responses side by side. The second response should be a formatted table tailored to your specific deadlines and schedule.

Both responses look similar?

Make sure you copied the full second prompt including all three deadlines and the table format instruction. If you only pasted part of it, Claude may still give a generic response.

Why do my two Claude responses look the same?

Practice three core prompting techniques

Every great prompt uses a combination of three techniques: providing context, specifying the output format, and giving examples of what good looks like. You'll practice all three in a single prompt.

Start a new conversation in Claude.
Paste this prompt that demonstrates all three techniques:

Context: I'm a university student studying computer science. I have a 10-minute presentation next week on "Why AI literacy matters for non-technical professionals."

Format: Give me an outline with exactly 3 main points. For each point, include:
- A one-sentence hook
- 2-3 supporting bullet points
- A transition sentence to the next point

Example of the tone I want:
- Good: "AI is already making decisions that affect your mortgage approval."
- Bad: "Artificial intelligence is a rapidly evolving field with many applications."

The outline should be conversational and use real-world examples that a business audience would relate to.

What are the three techniques here?

Context tells Claude who you are and what the task is for. This shapes the depth, tone, and relevance of the response.
Format tells Claude exactly how to structure the output. Instead of guessing, Claude follows your blueprint.
Examples show Claude the difference between what you want and what you do not want. This is the most powerful technique for controlling tone and style.

You should see a tightly structured outline that matches your format specification exactly. The tone should reflect your example style rather than defaulting to generic academic language.

Mix and match techniques

You do not need all three techniques in every prompt. For simple requests, context alone is enough. For style-sensitive tasks (emails, presentations, social posts), examples make the biggest difference.

Write a complex prompt using XML tags

For longer, more complex requests, plain text prompts can get confusing. XML tags let you clearly separate different parts of your prompt so Claude knows exactly where each piece of information starts and ends.

You'll write the same request two ways: first as a casual one-liner, then as a structured prompt with XML tags.

Start a new conversation in Claude.
First, send this casual one-liner to see what Claude produces:

Write me a professional email to my team about the new project timeline.

Notice how Claude has to guess the tone, the timeline details, and what your team needs to know. The result will be vague and likely unusable without heavy editing.

Now send this structured version using XML tags in the same conversation:

Write a professional email to my team based on the following details:

<context>
I'm a project manager at a software company. We just learned that our product launch is moving from March 15 to April 30 due to a vendor delay. The team has been working overtime and morale is low.
</context>

<requirements>
- Announce the new timeline clearly
- Acknowledge the team's hard work
- Explain the reason without blaming anyone
- Include 2-3 adjusted milestones
- End with something motivating
</requirements>

<tone>
Empathetic but confident. Not apologetic or overly casual. Similar to how a supportive manager would speak in a one-on-one, not a corporate memo.
</tone>

<format>
- Subject line included
- Under 250 words
- Use short paragraphs (2-3 sentences max)
</format>

Why do XML tags help?

XML tags like <context>, <requirements>, <tone>, and <format> create clear boundaries between different types of information. Claude can parse each section independently instead of trying to figure out where your context ends and your instructions begin.

This pattern scales to any complex task. The more moving parts your request has, the more XML tags help Claude keep everything organized.

Compare the two email outputs. The XML-tagged version should be a polished, ready-to-send email that matches your exact specifications for length, tone, and content.

Claude ignored some of your requirements?

Check that your XML tags are properly closed. Every <tag> needs a matching </tag>. If a tag is left open, Claude may misinterpret where one section ends and another begins.

Help me fix my XML-tagged prompt so Claude follows all my requirements

You now know how to write prompts that get focused, useful results from Claude. Next up, you'll create a dedicated Project that gives Claude persistent context about your life and work.

Create Your First Project

So far, every conversation you've had with Claude starts from scratch. Claude doesn't know who you are, what you're working on, or how you like to communicate.

Projects solve this problem by giving Claude a dedicated workspace with persistent context. You can set custom instructions that shape every response and upload reference files that Claude can draw on. In this step, you'll create a "Personal Productivity" project that turns Claude into a personalized assistant.

In this step, get ready to:

Create a new Project with custom instructions describing your role and preferences.
Upload a reference file to give Claude persistent context about you.
Start a conversation inside the project and confirm Claude uses your context.

Create a new project and write custom instructions

A project in Claude is like a dedicated workspace. It keeps your conversations, instructions, and uploaded files organized around a single topic. Custom instructions tell Claude how to behave in every conversation within that project.

In Claude, hover over the left side of the screen and click Projects.
Click + New Project in the upper right corner.
Enter Personal Productivity as the project name.
For the description, enter a brief summary like My personal productivity workspace for planning, decisions, and writing.
Click Create project to save it.

You should now see your project's main page with an empty knowledge base on the right side.

What is a Project in Claude?

Projects organize workspaces for ongoing conversations, resources, and related files. Each project has its own custom instructions and its own separate memory space. This means the context within each project stays focused and relevant.

On the free plan, you can create up to five projects.

On your project's main page, click Set project instructions.
Write instructions that describe your role, communication preferences, and the kind of tasks you'll bring to this project. Here's an example to customize:

## About me
- Role: [your role, e.g. "marketing manager at a startup" or "university student studying biology"]
- Goals: [your top 2-3 goals, e.g. "stay organized, hit deadlines, improve my writing"]

## Communication preferences
- Keep responses concise and actionable
- Use bullet points for lists
- Ask clarifying questions before tackling complex tasks
- Tone: friendly and direct

## Common tasks I'll bring here
- Weekly planning and task prioritization
- Decision-making frameworks
- Writing and editing drafts
- Brainstorming ideas

What do these instructions do?

Claude reads these instructions at the start of every conversation within this project. They shape the tone, format, and focus of every response without you having to repeat yourself each time.

The more specific you are about your preferences, the better Claude can tailor its responses. Replace the placeholder text in brackets with your actual details.

Replace the bracketed placeholder text with your own details.
Click Save instructions to apply your custom instructions.

Don't see the Set project instructions option?

Make sure you're on the project's main page, not inside a conversation. Navigate back by clicking the project name in the sidebar.

Help me find where to set custom instructions in my Claude project

Upload a reference file for persistent context

Custom instructions tell Claude how to behave. A reference file tells Claude who you are and what you're working on. Uploading a document gives Claude background knowledge it can draw on in every conversation.

Prepare a reference file from your computer. This could be your resume, a list of personal goals, a project brief, or a sample document that represents your work style.
On the right side of your project's main page, find the project knowledge base section.
Click the + button to add content to the project.
Select and upload your reference file.
Wait for Claude to process the file. You'll see it appear in the knowledge base list.

Choose a file that represents you

Your resume works great because it gives Claude your professional background, skills, and experience. A goals document works well because Claude can reference your priorities when helping you plan. Even a writing sample helps Claude match your style.

Anything you upload here will be used across all conversations within this project.

File not uploading?

Claude accepts common file formats like PDF, plain text, Word documents, and images. Make sure your file is not corrupted and is under the size limit shown in the upload dialog.

Help me troubleshoot my file upload to a Claude project

Start a conversation inside your project

Now it's time to test that your project setup is working. A conversation started inside a project automatically has access to your custom instructions and uploaded files.

From your project's main page, start a new conversation by clicking the chat area or the option to start a chat within the project.
Send a message that tests whether Claude knows your context. For example:

Based on what you know about me, suggest three things I should focus on this week to make progress toward my goals.

Review Claude's response. It should reference details from your custom instructions and uploaded file.

If Claude mentions your role, your goals, or specifics from your uploaded document, your project context is working correctly. The response should also follow your communication preferences (concise, bullet points, friendly tone).

Why does this feel different from a normal chat?

In a regular Claude conversation, you'd need to explain who you are and what you need every time. Inside a project, Claude already has that context loaded from your instructions and knowledge base. This is what makes projects powerful for ongoing work.

Claude's response seems generic?

Make sure you started the conversation from inside your project, not from the main Claude chat. Conversations started outside the project don't have access to your project's instructions or knowledge base.

Why isn't Claude using my project instructions in this conversation?

Your Personal Productivity project is now set up with custom instructions and a reference file. Every conversation you start inside this project will benefit from that persistent context. Next up, you'll use this project to build interactive artifacts that you can reuse every day.

Build Interactive Artifacts

You now have structured prompting skills that get dramatically better results from Claude. Time to put them to work building something you can actually use.

In this step, you will create interactive artifacts inside your project. Artifacts are standalone creations that Claude generates alongside its responses. They can be interactive HTML apps, documents, diagrams, and more. You will build two reusable tools without writing any code yourself.

In this step, get ready to:

Build an interactive weekly planner artifact with task management and priorities.
Iterate on your artifact with follow-up prompts and explore versioning.
Create a decision-making framework artifact with weighted criteria and recommendations.

Create your first artifact: an interactive weekly planner

Artifacts let you go beyond simple text responses. When you ask Claude to build something interactive, it generates a fully functional app that appears in a preview panel next to the conversation.

Switch back to your Personal Productivity project in the Claude sidebar.
Start a new conversation by clicking the compose or new chat area inside the project.
Send the following prompt to ask Claude to build your weekly planner:

Build me an interactive weekly planner as an HTML app. I want to be able to:
- Add tasks to specific days of the week
- Set priority levels (high, medium, low) for each task
- See my entire week at a glance in a clean layout
- Mark tasks as complete

Make it visually clean with a modern design. Use a card-based layout where each day is its own column.

Why this prompt works

This prompt uses the structured techniques you practiced in the last step. It specifies the output format (HTML app), lists concrete features, and describes the visual style. Claude recognizes this as an artifact request and generates a standalone interactive app.

Claude generates the artifact and displays it in a preview panel to the right of (or below) the conversation. You can interact with it immediately.

Click on the artifact preview to expand it and test the planner.
Try adding a few tasks to different days and setting different priority levels.
Mark one task as complete to confirm the interaction works.

What are artifacts?

Artifacts are standalone creations that live alongside your conversation. They can be single-page HTML websites, interactive React components, documents, code snippets, SVG images, or diagrams.

The key difference from a normal chat response is that artifacts are interactive and reusable. You can open them in a full window, share them, and come back to them later from the artifacts space in your sidebar.

Don't see an interactive preview?

Make sure you have Code execution and file creation enabled in Settings > Capabilities. Without this, Claude cannot generate interactive artifacts.

Also confirm you are using Claude on the web (claude.ai) or the desktop app. The artifacts tab is not currently available on mobile.

Help me troubleshoot why my artifact isn't appearing in Claude

Iterate on your artifact with follow-up prompts

Your first version is just the starting point. One of the most powerful aspects of artifacts is that you can refine them through conversation. Each follow-up prompt builds on the previous version while Claude maintains context about what you have built.

In the same conversation, send this follow-up prompt to add color coding:

Add color coding to the priority levels. Use red for high priority, orange for medium, and green for low. The color should appear as a left border or tag on each task card.

Claude updates the artifact in place. You should see the preview refresh with color-coded tasks.

Send another follow-up to change the layout and add a notes section:

Change the layout so weekdays are larger columns and weekend days are smaller. Also add a "Notes" section at the bottom of each day where I can jot quick reminders that aren't full tasks.

Your artifact now has three versions. You can scroll through previous versions to compare changes.

Look at the artifact panel for version indicators (small dots or arrows) that let you navigate between the original, color-coded, and final version.
Click between versions to see how each prompt changed the artifact.

How does versioning work?

Each time Claude updates an artifact based on your follow-up prompt, it creates a new version. You can navigate back to any previous version to compare. This means you can experiment boldly without losing earlier work.

If something breaks after a change, you can click Fix with Claude or describe the problem in plain language. You can also go back to a previous message and click Edit to create a new conversation branch and try a different approach.

Make it yours

Try one more iteration with something personal. Ask Claude to match your favorite color scheme, add a motivational quote section, or change the font. The goal is to make this a tool you will actually use every week.

Build a decision-making framework artifact

Your weekly planner helps you organize tasks. Now you will build a second artifact that helps you make better decisions. This tool will walk you through structured analysis when you are facing a tough choice.

Start a new conversation inside the same Personal Productivity project.
Send this prompt to create your decision-making framework:

Build me an interactive decision-making framework as an HTML app. It should guide me through making a tough decision step by step:

1. First, let me type in the decision I'm facing
2. Then let me add options I'm considering (at least 2)
3. Let me define criteria that matter to me (like cost, time, impact, risk)
4. Let me weight each criterion by importance (1-5 scale)
5. Let me score each option against each criterion
6. Calculate a weighted total and show a clear recommendation with a visual comparison

Make it feel like a guided wizard that walks me through each step one at a time, not all at once.

What makes this prompt effective?

This prompt gives Claude a numbered sequence of interactions to build. The "guided wizard" instruction tells Claude to reveal one step at a time rather than showing everything on a single screen. Describing the calculation logic (weighted totals) ensures the tool produces a meaningful recommendation.

Once the artifact appears, test it by entering a real decision you are facing.
Walk through all the steps: add options, define criteria, assign weights, and score each option.
Check that the final recommendation makes sense based on your scores.

If something feels off in the tool, use a follow-up prompt to fix it. You might ask Claude to adjust the scoring scale, change the visual layout, or add an explanation of why one option scored higher.

Framework not calculating correctly?

If the weighted scores seem wrong or the recommendation does not match what you expect, tell Claude what went wrong in plain language. For example: "The total scores don't seem to add up correctly" or "Option A should score higher than Option B based on my inputs."

You can also click Fix with Claude if you see an error message in the artifact preview.

Help me fix my decision-making artifact's scoring logic

You now have two interactive tools living inside your Personal Productivity project. Next up, you will learn how to generate real downloadable documents and analyze uploaded files.

Generate Real Documents and Files

Your Personal Productivity project already has interactive artifacts that live inside Claude. Now it's time to create files you can download and use outside of Claude.

Claude can generate real documents like Word files, PDFs, and spreadsheets using its code execution capability. This means you can go from a conversation to a polished deliverable in seconds.

In this step, get ready to:

Generate a downloadable document from a real task in your life.
Upload a file and have Claude analyze it.
Combine project context, structured prompts, and file creation into one polished deliverable.

Create a downloadable document from a real task

Artifacts are great for interactive tools, but sometimes you need a file you can email, print, or open in another app. Claude can create Excel spreadsheets, PowerPoint presentations, Word documents, and PDF files directly in your conversation.

What is code execution and file creation?

When you ask Claude to create a downloadable file, it writes and runs code behind the scenes to generate that file for you. You already enabled this capability in Step 1 under Settings > Capabilities.

You never see the code. You just get a download link for the finished file.

Think of a real task in your life that would benefit from a structured document. This could be a weekly meal plan, a personal budget spreadsheet, a project plan for something you're working on, or a study schedule.

In your Personal Productivity project conversation, send Claude a prompt like this one (replace the example details with your own real task):

Create a downloadable Word document with a weekly meal plan for one person.

Context:
- I cook 4 nights a week and eat out the other 3
- I'm trying to eat more vegetables and reduce processed food
- My budget is about $80/week for groceries
- I meal prep on Sundays

The document should include:
- A table with meals for each day (breakfast, lunch, dinner)
- A consolidated grocery list organized by store section
- Estimated cost per meal
- 2-3 meal prep instructions for Sunday

Format it professionally with headers, sections, and clean formatting.

What makes this prompt effective?

This prompt works because it gives Claude your personal context (cooking habits, budget, goals), specifies exactly what sections to include, and asks for a specific file format. The more detail you provide about your real situation, the more useful the output.

Wait for Claude to generate the document. You should see a file appear in the conversation with a download link.
Click the download link to save the file to your computer.
Open the downloaded file in Word, Google Docs, or your preferred app to confirm it looks correct.

Request a different format

If you'd prefer a spreadsheet for something like a budget, ask Claude to create an .xlsx file instead. For a presentation, ask for .pptx. Claude supports all of these formats.

Not seeing a download link?

Make sure you're using Claude on the web at claude.ai or on the desktop app. File creation works on web, desktop, and mobile.

If Claude gives you the content as text instead of a file, explicitly ask: "Please create this as a downloadable .docx file." Sometimes being direct about the file format helps.

If you're hitting usage limits on the free plan, wait a few minutes and try again.

Why isn't Claude generating a downloadable file for me?

Upload a file for Claude to analyze

Claude can also work in the other direction. Instead of creating files from scratch, you can upload an existing file and ask Claude to analyze, summarize, or transform it.

Find a file on your computer that you'd like Claude to analyze. Good options include:
In your project conversation, click the attachment icon (paperclip) at the bottom of the message input area.
Select your file from your computer.
Add a prompt that tells Claude what to do with the file:

I've uploaded [describe your file briefly]. Please:

1. Summarize the key points in 3-5 bullet points
2. Identify anything surprising or noteworthy
3. Create a clean, formatted summary I could share with someone else
4. Suggest 2-3 action items based on what you see

What does Claude do with uploaded files?

Claude reads the content of your file and can reference it throughout your conversation. For images, it uses vision to interpret what's shown. For documents and data, it extracts the text and structure.

Your uploaded file stays available for the entire conversation, so you can ask follow-up questions about it.

Send the message and review Claude's analysis.
Check that the summary accurately reflects your file's content. If Claude missed something important, ask a follow-up question to dig deeper.

Claude can't read your file?

Some file types may not be supported. Try converting your file to PDF or taking a screenshot of it and uploading the image instead.

If an image is blurry or has small text, Claude may struggle to read it. Try a higher-resolution version or type out the key content you want analyzed.

Help me get Claude to analyze my uploaded file

Combine everything into a polished deliverable

You've now used prompt engineering, project context, and file creation separately. The real power comes from combining all three in a single conversation to produce something genuinely useful.

In this final substep, you'll write one structured prompt that leverages your project's custom instructions and reference file to create a polished, downloadable document.

Start a new conversation inside your Personal Productivity project by clicking Start new chat within the project.
Write a structured prompt that combines context from your project with a specific file creation request. Here's an example using XML tags:

<task>
Create a professional project plan as a downloadable Word document.
</task>

<context>
Use what you know about me from this project's instructions and files.
The project I need to plan is: [describe a real upcoming project or goal]
Timeline: [your timeline, e.g., "next 4 weeks"]
</context>

<requirements>
- Executive summary (2-3 sentences explaining what this project achieves)
- Timeline with weekly milestones
- Task breakdown with owners (just me) and estimated hours
- Risk factors and mitigation strategies
- Success criteria: how I'll know it's done
</requirements>

<format>
- Professional Word document (.docx)
- Use headers, bullet points, and a table for the timeline
- Keep it to 2-3 pages maximum
- Tone: clear and actionable, not corporate jargon
</format>

Why does this work better inside your project?

Because you're inside the Personal Productivity project, Claude already knows your role, communication preferences, and has access to your uploaded reference file. The prompt says "use what you know about me" and Claude pulls from that project context automatically.

This is why you set up custom instructions earlier. Every conversation in this project benefits from that persistent context without you retyping it.

Replace the placeholder text in square brackets with your real information before sending.
Send the prompt and wait for Claude to generate the document.
Download the file and open it on your computer.
Review the document. If something needs adjusting, ask Claude in a follow-up message. For example: "Make the timeline table more detailed" or "Add a budget section."

Iterate until it's genuinely useful

The goal is a document you'd actually send to someone or use in your real life. Don't settle for a generic template. Keep refining with follow-up prompts until it reflects your actual situation.

Document feels too generic?

If the output reads like a template rather than something personal, your custom instructions may be too vague. Go back to your project settings and add more specific details about yourself.

You can also be more explicit in the prompt itself. Instead of "use what you know about me," reference specific details: "I work as a [your role] and my main constraint is [your constraint]."

How can I make Claude's documents more personalized to me?

You've just experienced one of Claude's most practical superpowers. You can now go from an idea to a finished, downloadable document in a single conversation. Next up, you'll configure memory and extended thinking to make Claude even more powerful across all your future conversations.

Configure Memory and Extended Thinking

You've generated real documents and built polished deliverables all within Claude. Your productivity system is taking shape.

Right now, every new conversation starts from scratch. Claude doesn't know your name, your role, or what matters to you. In this step, you'll teach Claude personal facts that persist across every conversation and test extended thinking for complex reasoning tasks.

In this step, get ready to:

Save personal facts (name, role, preferences, goals) to Claude's memory.
Test extended thinking on a complex multi-step problem.
Verify cross-conversation continuity by starting a fresh chat.

Teach Claude your personal facts

Claude's memory feature lets you tell it key facts about yourself that persist across conversations. Once saved, these facts inform every future response without you needing to repeat yourself.

Switch back to your conversation inside the Personal Productivity project.
Tell Claude your key personal facts by sending a message like this:

Remember these facts about me:
- My name is [your name]
- I'm a [your role, e.g., marketing manager, student, freelance designer]
- I prefer concise, actionable responses over long explanations
- My current goals are: [list 2-3 real goals, e.g., learn cloud computing, launch a side project, get promoted]

What does this do?

When you tell Claude to "remember" something, it updates your memory summary immediately. These facts will be available in your next conversation without you needing to re-send them.

Each project also has its own separate memory space. Facts saved here will influence responses within this project specifically.

Wait for Claude to confirm it has saved your facts.
Verify the save by navigating to Settings > Capabilities.
Click View and edit memory.
You should see the Manage memory modal displaying the personal facts you just shared (your name, role, preferences, and goals).

Add custom instructions to memory

You can also click the pencil icon in the lower left corner of the memory summary to add custom instructions directly. This is useful for fine-tuning how Claude remembers your preferences.

Don't see your facts in memory?

Make sure Memory is toggled on in Settings > Capabilities. If it was off, toggle it on and resend your message asking Claude to remember your facts.

If Claude acknowledged the save but the memory modal is empty, refresh the page and check again. Edits apply immediately to your next conversation.

Why isn't Claude saving my personal facts to memory?

Test extended thinking with a complex problem

Extended thinking lets Claude reason through difficult, multi-step problems before responding. Instead of jumping straight to an answer, Claude takes time to consider different angles, weigh tradeoffs, and build a more thorough response.

Start a new conversation inside your Personal Productivity project (click the compose button or New chat).
Give Claude a complex, multi-step problem that requires deep reasoning. Send a message like this:

I need your help thinking through a major decision. I'm considering [choose one that's real to you: changing careers to tech, starting a side business, moving to a new city, or designing a 6-month learning path to become a data analyst].

Please think deeply about this and give me:
1. A framework for evaluating this decision
2. The key risks and how to mitigate each one
3. A realistic timeline with milestones
4. What I should do in the first 2 weeks to test this path

What happens during extended thinking?

When Claude encounters a complex problem, it uses extended thinking to reason step-by-step before generating a response. You may see a brief "thinking" indicator while Claude works through the problem.

The response quality improves significantly for multi-step reasoning tasks like planning, analysis, and decision-making. Compare this structured response to what you'd get from a one-sentence question.

Read through Claude's response. Notice how the answer is structured, addresses multiple dimensions of the problem, and provides actionable next steps.
Follow up with a probing question to test deeper reasoning. For example:

What's the biggest blind spot in this plan? What am I most likely to underestimate?

Claude should demonstrate self-reflection on its own response, identifying gaps or assumptions it made in the original analysis.

Response seems short or shallow?

Try making your prompt more specific. The more context and constraints you give Claude, the more it has to reason about. Add details about your situation, timeline, budget, or constraints.

If you're hitting usage limits on the free tier, wait a few minutes and try again. Extended thinking uses more capacity per response.

How can I get Claude to think more deeply about my problem?

Verify cross-conversation memory

The real power of memory is continuity. Let's confirm that Claude remembers your preferences in a brand new conversation without you reminding it.

Start a completely new conversation (outside of any project). Click the compose button in the sidebar.
Send a simple message that relies on Claude knowing who you are:

Based on what you know about me, suggest three things I should focus on this week.

Check that Claude's response references the facts you saved earlier (your name, role, or goals).
If Claude responds with generic suggestions without referencing your stored preferences, it means memory isn't carrying over. Go back to Settings > Capabilities and confirm Memory is toggled on.

Why does this matter?

Without memory, you'd have to re-explain who you are every single time you open Claude. Memory transforms Claude from a generic chatbot into a personalized assistant that gets better the more you use it.

Your memory summary is updated automatically every 24 hours based on your conversations. You can also update it instantly by telling Claude what to remember, just like you did earlier.

Claude doesn't remember your facts?

If you started the new conversation inside a different project, its memory space is separate. Try starting a standalone conversation (not inside any project) to test the global memory.

If the global memory is still empty, navigate to Settings > Capabilities > View and edit memory and manually add your facts using the pencil icon.

Help me get Claude's memory working across conversations

Secret mission

Build an AI-Powered Personal Coach

You've built tools that organize your week and help you make decisions. But what if Claude could actively coach you through a challenge, asking smart questions and adapting its advice based on your answers? In this mission, you'll build an AI-powered artifact that acts as your personal coach.

Clean Up Your Resources

Decide whether to keep your resources running, pause them to come back later, or delete them entirely. This project runs entirely on Claude's free tier, so there are no ongoing costs.

Resources you used:

Claude Memory entries (your name, role, preferences, and goals).
Personal Productivity Project (custom instructions and uploaded reference file).
Artifacts (weekly planner, decision-making framework, personal coach, and generated document).
Downloaded files on your computer (Word or PDF documents).

Keep everything running

No action needed. Choose this if you're still actively using your productivity system and want Claude to keep remembering your preferences.

Your memory entries, project, and artifacts all stay intact on the free plan.
You can return to your Personal Productivity Project anytime from the Claude sidebar.
Your artifacts remain accessible in the Artifacts tab in the sidebar.

Pause - I'll come back to this later

Since Claude is a free web service, there is nothing to shut down or pause. Your resources stay available at no cost whenever you return.

Your memory, project, and artifacts will persist in your account.
If you want to reduce personal data exposure while you're away, you can clear your memory entries. Navigate to Settings > Capabilities > View and edit memory and remove any entries you'd prefer not to store.

Delete - I don't want to use this again

Remove all project resources and personal data from Claude. Follow each step below.

Clear your memory entries. Navigate to Settings > Capabilities > View and edit memory. Delete each memory entry that Claude saved about you (your name, role, preferences, and goals).
Delete your Personal Productivity Project. In the Claude sidebar, find the Personal Productivity project. Open the project settings and delete it. This removes the custom instructions and uploaded reference file.
Delete your artifacts. Open the Artifacts tab in the sidebar. Delete the weekly planner, decision-making framework, personal coach, and any other artifacts you created during this project.
Delete downloaded files from your computer. Find and delete any Word or PDF documents you downloaded during Step 5. Check your Downloads folder in Finder (macOS) or File Explorer (Windows).

Nice Work!

You did it. You've gone from zero to Claude power user with a fully personalized productivity system that you can use every single day.

Here's what you accomplished:

Mastered prompt engineering techniques (context, format, examples, XML tags) that produce dramatically better results from any AI conversation.
Created a personalized Project with custom instructions and uploaded reference files, giving Claude persistent context about your work and preferences.
Built interactive artifacts (a weekly planner and a decision-making framework) and generated real downloadable documents, all without writing a single line of code yourself.
Completed the Secret Mission: built an AI-powered coaching artifact that asks questions, adapts to your responses, and delivers personalized recommendations.

Ready to quiz yourself?