How Claude Projects work.
Why this mattersA Claude Project is the difference between re-explaining your business every time and working with an assistant that already knows it. Understanding how it works is what makes every prompt in this course actually land.
Read this once. Sit with it before you answer.
Why does AI feel like it forgets everything you told it — and what would change if it did not?
Where you stand right now.
By the end of this lesson, you will be able to:
- 1Explain how a Claude Project differs from a regular Claude chat.
- 2Describe the three parts of a project and what each one does.
- 3Explain how the course prompt files work and how you will run them by number.
The whole lesson in five points.
- 01A regular Claude chat starts cold every time. A Claude Project does not.
- 02A project has three parts: knowledge files, Project Instructions, and the conversations you have inside it.
- 03Knowledge files are your business reference library. Claude reads them on every task in the project.
- 04The course gives you prompt files. You load them into your project and run them by number, like run Prompt 2.
- 05Prompts only work inside the project, because that is where your business information lives.
Ten minutes on how the tool actually works.
Most people who try AI for their business quit for one reason: they feel like they are explaining themselves over and over to something that never remembers.
A Claude Project fixes that. Once you understand how it works, every prompt in this course makes sense, and you stop fighting the tool.
A regular chat versus a project.
Start with a regular Claude chat, the kind you may have used before. You open it, you explain your business, you get an answer, you close it. Tomorrow you open a new chat and Claude knows nothing. You explain your business again. Every conversation starts cold.
A Claude Project changes that. Think of a regular chat as talking to a sharp assistant who has amnesia between every meeting. A project is the same assistant, but now they have a folder on their desk with everything about your business in it, and they read that folder before every task.
Same intelligence — now informed. That folder is the whole point of a project, and it is what you are about to build.
The three parts of a project.
Knowledge files
Documents you upload to the project — your Quick Info, your reviews, your competitor notes. Claude reads all of them on every task inside the project, automatically. This is the folder on the desk. It does not expire, and you can update it whenever your business changes.
Project Instructions
A short, standing set of rules that tells Claude how to behave in this project, every time. Your brand voice, your customer language, your rule against em dashes. You write it once and it applies to everything.
The conversations
The chat area where you actually run prompts and do work. You can start as many separate conversations inside the project as you want, and they all share the same knowledge files and the same instructions.
Knowledge files are what Claude knows. Project Instructions are how Claude behaves. The conversation is where the work happens.
How the course prompts work.
This course runs on prompts, and there are fifteen of them. You are not going to copy and paste long prompts one at a time. The course gives you the prompts as files — named Prompt 1, Prompt 2, and so on through Prompt 15.
You load those prompt files into your project once, alongside your business files. From then on, running a prompt is simple. You open a conversation in your project and you type run Prompt 2. Claude finds the Prompt 2 file, reads it, reads your Quick Info, and produces the output.
Each prompt file begins with one line that tells Claude the file is a prompt to execute, not a document to summarize. You do not need to understand that line. You just need to know that the prompts live in your project and you run them by number.
Why prompts only work inside the project — and what to do when something looks wrong.
One rule matters more than any other. Prompts only work inside your project. If you open a plain Claude chat with no project and try to run a course prompt, Claude has no Quick Info, no reviews, nothing about your business. The output will be generic and close to useless.
So every time you run a prompt: open Claude, open your project, start the conversation there.
Two quick troubleshooting notes. If a prompt gives you generic output that could belong to any business, you are either not inside your project or your knowledge files did not upload. Check both. And if you ask Claude to run a prompt and it describes the prompt back to you instead of running it, simply reply: run it fully now, every section.
One practical thing. On the free Claude plan there is a limit to how many messages you can send in a stretch. If you run several prompts back to back and hit the limit, wait a few hours or consider upgrading. For one module per sitting, the free plan is enough.
A regular chat forgets you. A project remembers, because your business lives in its files.
You now know the three parts — knowledge files, instructions, and conversations — and you know the prompts are files you run by number. Next, you build your own project and load it.
The vocabulary that follows you.
- Knowledge files
- The documents you upload to a Claude Project. Claude can read them automatically on every task inside that project.
- Project Instructions
- The short, standing set of rules inside a project that tells Claude how to behave every time: your voice, your context, your default rules.
- Prompt file
- A course-provided file containing one runnable prompt, stored in your project and run by its number, such as run Prompt 2.
Two quick things before the next lesson.
First, download the prompt files provided with this course, so they are ready to upload when you build your project. Second, open claude.ai and confirm you can see Projects in the left sidebar. That is all. The next lesson is the build.