wiki/knowledge/ai-tools/claude-team-project-workflow.md Layer 2 article 646 words Updated: 2026-04-05
↓ MD ↓ PDF
ai-tools claude-ai workflow content-generation instructional-design collaboration

Claude Team Project Workflow for Content Generation

Overview

When developing training content collaboratively, Claude AI team projects provide a shared knowledge base that multiple contributors can query for content generation, refinement, and question-answering. This article documents how to set up and use Claude Team projects effectively in a training development context, based on practices established during the [1] SOAR training program build.

Account Structure

Claude distinguishes between personal and team accounts. The team account (identified by a building icon in the lower-left of the interface) is the company account under Asymmetric. Use it for all work-related queries.

Switching between accounts is done via the icon in the lower-left corner of the Claude interface.

Projects vs. Chats

Claude organizes work into two areas:

Area Icon Use
Chats Speech bubble General, one-off queries not tied to a specific project
Projects Filing drawer Topic-specific work with uploaded source documents

Only use a project chat when your query is directly related to that project's content. For unrelated questions, use a general chat to keep projects clean and focused.

Accessing a Shared Project

  1. Log in to Claude using your team account (asymmetric.pro email)
  2. Click the filing drawer icon on the left sidebar to open Projects
  3. The shared project will appear once the project owner has added you as a member and shared the project
  4. If the project doesn't appear immediately, refresh — large document sets can be slow to load

Note: Invitation links expire quickly. If you receive an "Invalid Invite Link" error, ask the project owner to remove and re-add you as a member.

How to Query a Project for Content Work

When working through client feedback or content gaps, use the project chat to ask targeted questions. A useful pattern:

  1. Reference the specific module or lesson you're working on
  2. Paste the relevant content block if needed for context
  3. State the client's comment or the gap you're trying to fill
  4. Ask for a recommendation

Example prompt structure:

"I'm working on Module 3, the competitive landscape section. The client said [X]. Based on the project materials, what would you recommend we change or add?"

If the project returns a useful answer, implement it. If the answer doesn't make sense or the content isn't in the project, flag it for manual editing or regeneration.

Decision Framework: When to Use Claude vs. Manual Edits

Not every content issue requires AI assistance. Use this triage approach:

Situation Action
Clear, simple fix (typo, ordering, formatting) Edit directly
Ambiguous client feedback Ask the client for clarification before editing
Content gap or rewrite needed Query the Claude project; implement if output is sound
Content missing from project entirely Regenerate via Claude; note what was regenerated
Client feedback may be outdated Check whether the content was already updated before acting

Practical Notes from the Agility Project