Module 10 · Lesson 05

Claude.ai Chat — Projects, Memory, and Team Spaces

Reading time: 15 minutes Track: Claude Fluency for Teams · Required for all learners

The organizational layer of Claude.ai

Most people start using Claude.ai as a series of one-off conversations. This works, but you lose all context between sessions and end up re-explaining your situation repeatedly. Claude.ai's organizational features — Projects, Memory, and Team features — eliminate that overhead.

Projects

A Project is a persistent context container. You create a Project for a recurring workstream — a product launch, a client engagement, a research area — and everything within that Project remembers the context you've established.

What a Project gives you:

• Project instructions: Persistent system prompt for all conversations in the Project. "You're helping with our Q3 product launch. Here's the product overview, our target audience, and our key messaging..." — written once, active forever.
• Shared context: Documents, notes, and files you add to the Project are available in all Project conversations
• Organized history: All conversations in a Project are grouped together

When to create a Project:

• Any workstream you return to more than once per week
• Any context that would take >2 minutes to re-explain
• Any task where consistency across sessions matters (tone, decisions made, constraints established)

Setting up a Project:

1. In Claude.ai sidebar, click "+ New Project"
2. Name it clearly (it's for your benefit)
3. Click "Project instructions" and write your persistent context
4. Add any relevant documents

What to put in Project instructions:

Context: I'm a product manager at a 50-person B2B SaaS company.
This project is for [specific workstream].

Background I always want you to have:
- [Key context about the situation]
- [Constraints or preferences]
- [Any style or format preferences]

When responding, always [any behavioral preference].

Memory

Claude.ai can maintain a personal memory — facts about you that persist across all conversations. This is separate from Projects; it applies globally.

Useful things to add to memory:

• Your role and what you work on
• Your communication style preferences ("I prefer direct answers, skip preamble")
• Your technical environment ("I primarily use Python and work on macOS")
• Standing preferences ("Always suggest tests when I share code")

Memory can be managed under Settings → Memory. You can view, edit, and delete specific memories.

Team features (Claude.ai Teams/Enterprise)

Team spaces: Shared Projects that multiple team members can access. Useful for:

• A shared research Project where multiple people contribute
• A standard instructions template your whole team uses
• Onboarding resources new hires can reference

Admin controls (Enterprise): Centralized policy, usage analytics, member management, SSO integration.

The practical setup for a team

For a team getting started, this setup takes about 30 minutes and pays off every week:

1. Each person creates a personal memory entry with their role and key context
2. Create Projects for your 3-5 most recurring workstreams
3. Write Project instructions for each (aim for 100-300 words of context)
4. Add any documents the Project should always reference
5. For shared work: create a Team space and share the Project

The payoff: conversations in Projects start at context level 7 instead of level 1.