- AI Business Insights
- Posts
- The Only Claude Cowork Setup Guide You Need
The Only Claude Cowork Setup Guide You Need
The setup everyone skips
Most people use Claude like a search engine. They type a question, get an answer, and start the next session with zero memory of what came before. I was doing the exact same thing until I came across this post from a LinkedIn creator who laid out the complete Claude Cowork setup, step by step, folder by folder, prompt by prompt.
The result? Prompts that are 10 words long. Outputs that actually sound like you. A workflow where Claude asks you questions instead of guessing. Here’s the full breakdown.
How Jennifer Anniston’s LolaVie brand grew sales 40% with CTV ads
The DTC beauty category is crowded. To break through, Jennifer Anniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.
Step 1: Download the App and Lock Your Settings
The author starts with the basics: download the Claude desktop app, open the Cowork tab, select the “Opus 4.6” model, and turn on Extended Thinking. Those three settings stay fixed. The creator is emphatic about this: never change them. This specific combination is the foundation for everything that follows, and deviating from it will undermine the whole system.
Step 2: Build the Right Folder Structure
Create one master folder on your laptop called “Claude-Work.” Inside it, add four subfolders: About Me, Project, Template, and Outputs. Then point the Cowork tab to this master folder. This is how Claude “reads” you before every task, giving it real context about who you are and how you work before a single word of your prompt is typed.
Step 3: Write Your Three Context Files
This is the step most people skip, and it’s the one that matters most. The original poster recommends creating three specific files inside the About Me subfolder:
about-me: what you actually do day-to-day, not a resume, but a real description of your work and responsibilities
my-voice: your tone, phrases you hate, and three solid examples of your writing style
my-rules: instructions like “ask before executing,” “show a plan,” and “never delete without approval”
These files transform Claude from a generic assistant into something that sounds like you. Without them, outputs will always feel off-brand and hollow, no matter how carefully you write your prompts.
Stop typing prompts. Start talking.
You think 4x faster than you type. So why are you typing prompts?
Wispr Flow turns your voice into ready-to-paste text inside any AI tool. Speak naturally - include "um"s, tangents, half-finished thoughts - and Flow cleans everything up. You get polished, detailed prompts without touching a keyboard.
Developers use Flow to give coding agents the context they actually need. Researchers use it to describe experiments in full detail. Everyone uses it to stop bottlenecking their AI workflows.
89% of messages sent with zero edits. Millions of users worldwide. Available on Mac, Windows, iPhone, and now Android (free and unlimited on Android during launch).
*Ad
Step 4: Set Global Instructions Once, Forever
Go to Settings, then Cowork, then Edit Global Instructions. The expert recommends pasting this exact text:
“I’m [Name], [Role]. Read my files before every task. Ask questions before executing. Show a plan. Never delete without approval.”
Set it once. It runs automatically at the start of every session. You’ll never need to type it again. This is what makes short prompts possible because the context is already loaded before you say a single word. The original poster calls this one of the highest-leverage moves in the whole setup.
Step 5: Stop Prompting, Start Answering
Here’s the shift that makes this system genuinely different. Instead of writing long, detailed prompts from scratch, the creator uses this exact template for every task:
“I want to [TASK]. Read all files first. Ask me questions using AskUserQuestion before you execute. If something is off, generate new questions. Do not guess.”
When you prompt this way, Claude generates a clickable question form instead of a wall of text. You select answers rather than typing paragraphs of context. The original poster calls this “stop prompting, start answering questions,” and it completely changes how the interaction feels. You’re steering, not dictating, and the outputs reflect that difference immediately.
Step 6: Install Your First Plugins
Navigate to Customize, then Browse Plugins, then Install. Type “/” to see available slash commands. The creator suggests starting with just two plugins based on your actual work: marketing, data, legal, or sales are the recommended starting points. Keep it focused at first so you actually learn how the slash commands work before adding more. This plugin layer is, as the post’s author notes, one of the core reasons so much traditional software value has been disrupted recently.
Step 7: Connect Your Real Tools
Go to Settings, then Connectors, then Add. Claude can connect to Slack, Google Drive, Notion, Figma, and 50+ other tools. Once connected, Claude reads your tools mid-conversation. No copy-pasting files across tabs. No switching contexts. The original poster points out that this works on all plans at no extra cost, which makes it one of the most underused features in the whole app.
The Don’ts
The creator’s list of habits that break this system is worth keeping close:
Writing long prompts for every task instead of relying on your context files
Skipping the folder structure and wondering why outputs feel generic
Dumping 50 random files into the folder without any structure or labels
Using the Chat tab when you actually need the Cowork tab
Sending deliverables without reviewing them first
Forgetting to turn on Extended Thinking
Expecting Claude to know your preferences without context files loaded
Quitting the desktop app mid-session
The Do’s
Always select Opus 4.6 and Extended Thinking before starting
Always point Cowork to your master folder before any task
Always say “ask me questions first” before Claude executes anything
Keep your three context files updated as your role and projects evolve
Let Claude ask its questions rather than pre-answering everything in the prompt
Always review outputs before they go anywhere
Connect your real tools through Connectors so Claude works in your actual environment
Invest the 30 minutes in setup upfront, not in crafting longer prompts later
The core idea here is this: instead of getting better at prompting, get better at context. When Claude knows who you are, what you sound like, and what your rules are, a 10-word prompt can produce polished, on-brand work every single time.
Check out the full LinkedIn post to see how the original poster structures their own context files and to grab their exact setup for yourself.
Email Still Wins. Here's How to Use It Better.
59% of Americans say most marketing emails offer no real value. That's not a threat, it's an opening. Get the AI-powered playbook for building email campaigns that actually convert.
Inside you'll discover:
How top brands achieve 3,600% ROI from email marketing
AI personalization techniques that drive 82% higher conversion rates
Tactics that have delivered 30% better open rates and 50% higher clickthroughs
How to build sequences for every stage of the customer journey, from welcome to re-engagement
Download your free AI-powered email marketing playbook today.
*Ad



