Module 2
How to give Claude the context it needs to do your best work - and keep it across every conversation.
By default, every new conversation with Claude begins with no memory of who you are, what you are working on, or what decisions you have already made. That is fine for a quick question. It is not fine for the kind of sustained, context-dependent work executives actually do.
Think about what that means in practice. You ask Claude to help draft a strategy memo. You spend the first ten minutes explaining your organization, your audience, your constraints, your tone. You get something useful. Next week, same problem, different angle - and you start over. Claude does not remember the memo. It does not remember your organization. It does not remember anything.
Starting a new Claude conversation without a Project is like hiring a skilled consultant each morning who has no memory of your industry, your organization, or anything you discussed the day before. You spend the first twenty minutes re-explaining context that should already be understood. Projects eliminate that tax.
A Claude Project is a persistent workspace. Four components work together to give Claude the foundation it needs before you type a single word.
Written once, active in every conversation. Who you are, what this project is, how you want Claude to think with you. The difference between a generic assistant and one that already knows your context.
Documents Claude reads before responding. Strategy papers, brand guidelines, previous analyses, research. Claude can reference these directly - you do not have to paste them into every conversation.
Claude builds understanding over time from your conversations within the project - automatically. Decisions made, directions taken, conclusions reached. Review your project memory periodically: Claude decides what to retain, and it is worth confirming the most important context is captured accurately.
All conversations within the project live in one place, organized and searchable. You can return to any thread, pick up where you left off, or trace how a decision evolved over multiple sessions.
In today's session, you looked at the AI Coaching project - the workspace where this entire program was built. Here is what you were actually looking at.
Claude knows the program philosophy, the participant profiles, the teaching approach, and the brand standards - before a single question is asked.
Handoff documents, session guides, curriculum gap analysis, brand style guide, business plan. Claude reads these and references them directly in every conversation.
Building a 9-module curriculum. Cohort development. Module 1 review. Module 2 preparation. Each conversation built on the last.
Every page you have read, every session structure you have followed, every decision about sequencing these nine weeks was built inside this project.
The curriculum you are reading, the session Jay is running today, the handoff documents that carry context from one week to the next - all of it was built in conversation with Claude inside that project. You are looking at the output of the tool you are about to learn to use.
Not every task needs a Project. The question is whether Claude needs to know something about you, your work, or your context in order to give you something genuinely useful - and whether you will need that same context more than once.
Ask yourself: if Claude already knew everything about this problem, how much better would the answer be? If the answer is "significantly" - a Project is worth the fifteen minutes it takes to set up.
If a conversation that felt self-contained starts growing into something ongoing, you can move it into a project at any time. Open the conversation, click the project menu, and reassign it. The full history moves with it.
Before Module 3, create one Claude Project around a real problem in your work. It does not need to be perfect. It needs to be real. The only way to understand what a Project changes is to experience the difference it makes on something you actually care about.
Choose something you return to regularly - a strategic initiative, a persistent challenge, an ongoing relationship you manage. It should be something where Claude having context about your role, your organization, or your goals would change the quality of what you get back.
A paragraph or two. Who you are, what this project is for, and how you want Claude to approach it with you. Do not overthink this - it is a brief, not a contract. You can refine it after your first few conversations reveal what Claude needs to know.
A document that captures something Claude should always know. A strategy memo, a set of guidelines, a previous analysis, a brief. Start with whatever single document would give Claude the most useful foundation for this problem.
Use the project for actual work this week - not a test prompt, not a demo. Bring a real question from your real work and see what changes when Claude already knows the context. Then compare it to what you would have gotten starting from scratch.
Come to Module 3 ready to show what you built. Share what problem you chose, what you put in the instructions, and what was different about the conversation you had. Both the wins and the moments that did not work yet are worth discussing.
The feature descriptions here reflect Claude Projects as of early 2026. Anthropic updates Claude regularly, and specific interface details may change. If something described here does not match what you see in Claude, check the support documentation linked above. The underlying concept - giving Claude persistent context - will not change even as the interface evolves.