How to Use Claude Projects: Build a Brand-Specific AI Assistant with Custom Instructions and Knowledge Files

How to Use Claude Projects: Build a Brand-Specific AI Assistant with Custom Instructions and Knowledge Files

Claude Projects is one of the most powerful features available to Claude Pro, Team, and Enterprise users. It lets you create persistent workspaces where you define custom instructions and upload knowledge files, effectively transforming Claude into a specialized AI assistant tailored to your brand, workflow, or domain. This guide walks you through the complete setup process with real examples.

What Are Claude Projects?

A Claude Project is a dedicated workspace within Claude that maintains context across conversations. Unlike standard chats that start from zero every time, Projects allow you to:

  • Set persistent custom instructions that shape every response- Upload knowledge files (up to 200K tokens of context) that Claude references automatically- Share the project across team members (on Team and Enterprise plans)- Create multiple specialized assistants without repeating setup

Step-by-Step: Creating Your Brand AI Assistant

Step 1: Create a New Project

  • Open claude.ai and log in with your Pro or Team account.- In the left sidebar, click Projects.- Click Create Project and give it a descriptive name, e.g., Acme Brand Assistant.- Add an optional description: “AI assistant for Acme Corp content, support, and internal communications.”

Step 2: Write Custom Instructions

Click the Set custom instructions field inside your project. This is where you define Claude’s persona, tone, rules, and behavioral constraints. Here is an example instruction set: You are the official AI assistant for Acme Corp.

Brand Voice

  • Tone: Professional yet approachable. Never use slang.
  • Always refer to the company as “Acme Corp” (never “Acme” alone).
  • Use active voice. Keep sentences under 25 words when possible.

Response Rules

  • Always include a call-to-action when answering customer-facing queries.
  • For pricing questions, direct users to acmecorp.com/pricing.
  • Never disclose internal roadmap details or unannounced features.
  • When unsure, say: “Let me connect you with our team for the most accurate answer.”

Formatting

  • Use headers (H2, H3) for long responses.
  • Use bullet points for lists of 3+ items.
  • Include relevant links from the knowledge base when available.

Audience

  • Primary: Small business owners (non-technical)
  • Secondary: Enterprise procurement teams

Step 3: Upload Knowledge Files

Click **Add content** in the project knowledge section. You can upload multiple file types:

File TypeBest ForMax Size
PDFBrand guidelines, whitepapers, product manuals~30MB per file
TXT / MDFAQs, style guides, structured knowledge bases~30MB per file
CSVProduct catalogs, pricing tables, feature matrices~30MB per file
Code files (.py, .js, etc.)API documentation, code standards~30MB per file
**Recommended knowledge files for a brand assistant:** - brand-guidelines.pdf — Logo usage, color palette, typography, voice and tone- product-faq.md — Frequently asked questions and approved answers- pricing-sheet.csv — Current pricing tiers, features per plan- support-scripts.txt — Common support scenarios and resolution steps- competitor-comparison.md — Positioning against competitors (internal use) ### Step 4: Test and Iterate Start a new conversation within the project and test with real-world queries: You: "What makes Acme Corp different from CompetitorX?" You: "Write a follow-up email for a customer who asked about enterprise pricing." You: "Summarize our refund policy in 3 bullet points."

Review the outputs against your brand standards. If Claude deviates, refine your custom instructions. Common adjustments include: - Adding explicit **"Do not"** rules for edge cases- Providing example outputs in the instructions (few-shot prompting)- Restructuring knowledge files for better retrieval ### Step 5: Integrate via the API (Optional) For developers who want to use Projects programmatically, you can reference project context through the Claude API: curl https://api.anthropic.com/v1/messages \ -H "Content-Type: application/json" \ -H "x-api-key: YOUR_API_KEY" \ -H "anthropic-version: 2023-06-01" \ -d '{ "model": "claude-sonnet-4-20250514", "max_tokens": 1024, "system": "You are the Acme Corp brand assistant. Follow all brand guidelines strictly.", "messages": [ { "role": "user", "content": "Draft a product announcement for our new analytics dashboard." } ] }'

For persistent context equivalent to Projects via API, use the system prompt combined with document content passed in the message context.

Pro Tips for Power Users

  • Layer your instructions: Put universal rules (tone, formatting) at the top and domain-specific rules (product details, compliance) lower. Claude weighs earlier instructions more heavily in ambiguous situations.- Use structured knowledge files: Markdown with clear headers outperforms unstructured text. Use ## Section Name headings in your .md files so Claude can locate information faster.- Create multiple projects per function: Instead of one mega-project, build separate projects for Sales, Support, Content, and Engineering. This keeps context focused and reduces noise.- Version your instructions: Keep a changelog at the bottom of your custom instructions. Example: v2.3 — 2026-03-15: Added compliance language for EU customers.- Test adversarially: Ask Claude to write content that would violate your guidelines and verify it refuses or self-corrects. This reveals gaps in your instruction set.

Troubleshooting Common Issues

ProblemCauseSolution
Claude ignores uploaded knowledge filesFile content is too unstructured or the query doesn't trigger retrievalAdd explicit headers in your files. In instructions, add: "Always check the uploaded knowledge base before answering product questions."
Responses don't match brand toneInstructions are too vagueAdd 2-3 example responses directly in your custom instructions as few-shot examples
Claude contradicts pricing or feature infoOutdated knowledge files or conflicting information across filesAudit all uploaded files for consistency. Remove outdated versions and re-upload current ones
Context window exceeded warningToo many or too large knowledge files uploadedConsolidate files. Remove redundant content. Prioritize the most-referenced documents
Team members see different behaviorIndividual chats outside the projectEnsure all team members start conversations from within the shared project, not from the main chat
## Frequently Asked Questions

How many knowledge files can I upload to a single Claude Project?

You can upload multiple files to a Claude Project, with a combined limit of approximately 200,000 tokens of context (roughly 500 pages of text). The practical limit depends on file sizes, but most brand assistants work well with 5-10 focused documents rather than dozens of large files. Prioritize quality and relevance over quantity.

Can I share a Claude Project with my team, and will they see the same custom instructions?

Yes. On Claude Team and Enterprise plans, Projects can be shared with team members. All participants see and use the same custom instructions and knowledge files, ensuring consistent AI behavior across your organization. Individual team members cannot override the project-level instructions, which maintains brand consistency.

What is the difference between custom instructions in Projects and the system prompt in the API?

Claude Projects custom instructions are set through the claude.ai web interface and persist across all conversations within that project. The API system prompt serves a similar purpose but is set programmatically per request. Projects are ideal for non-technical users who need a persistent, no-code setup. The API system prompt offers more flexibility for developers building custom applications, but requires you to manage context and document uploads manually in each request.

Explore More Tools

Grok Best Practices for Academic Research and Literature Discovery: Leveraging X/Twitter for Scholarly Intelligence Best Practices Grok Best Practices for Content Strategy: Identify Trending Topics Before They Peak and Create Content That Captures Demand Best Practices Grok Case Study: How a DTC Beauty Brand Used Real-Time Social Listening to Save Their Product Launch Case Study Grok Case Study: How a Pharma Company Tracked Patient Sentiment During a Drug Launch and Caught a Safety Signal 48 Hours Before the FDA Case Study Grok Case Study: How a Disaster Relief Nonprofit Used Real-Time X/Twitter Monitoring to Coordinate Emergency Response 3x Faster Case Study Grok Case Study: How a Political Campaign Used X/Twitter Sentiment Analysis to Reshape Messaging and Win a Swing District Case Study How to Use Grok for Competitive Intelligence: Track Product Launches, Pricing Changes, and Market Positioning in Real Time How-To Grok vs Perplexity vs ChatGPT Search for Real-Time Information: Which AI Search Tool Is Most Accurate in 2026? Comparison How to Use Grok for Crisis Communication Monitoring: Detect, Assess, and Respond to PR Emergencies in Real Time How-To How to Use Grok for Product Improvement: Extract Customer Feedback Signals from X/Twitter That Your Support Team Misses How-To How to Use Grok for Conference Live Monitoring: Extract Event Insights and Identify Networking Opportunities in Real Time How-To How to Use Grok for Influencer Marketing: Discover, Vet, and Track Influencer Partnerships Using Real X/Twitter Data How-To How to Use Grok for Job Market Analysis: Track Industry Hiring Trends, Layoff Signals, and Salary Discussions on X/Twitter How-To How to Use Grok for Investor Relations: Track Earnings Sentiment, Analyst Reactions, and Shareholder Concerns in Real Time How-To How to Use Grok for Recruitment and Talent Intelligence: Identifying Hiring Signals from X/Twitter Data How-To How to Use Grok for Startup Fundraising Intelligence: Track Investor Sentiment, VC Activity, and Funding Trends on X/Twitter How-To How to Use Grok for Regulatory Compliance Monitoring: Real-Time Policy Tracking Across Industries How-To NotebookLM Best Practices for Financial Analysts: Due Diligence, Investment Research & Risk Factor Analysis Across SEC Filings Best Practices NotebookLM Best Practices for Teachers: Build Curriculum-Aligned Lesson Plans, Study Guides, and Assessment Materials from Your Own Resources Best Practices NotebookLM Case Study: How an Insurance Company Built a Claims Processing Training System That Cut Errors by 35% Case Study