Workspace Chats

What are Workspace Chats?

Workspace chats are a powerful feature that allows you to interact directly with the documents and knowledge stored in your workspace using advanced AI technologies. It's like having a smart assistant that can:

• Read and understand all documents in your workspace
• Answer questions based on your specific document collection
• Provide insights and summaries
• Help you explore and retrieve information

How Workspace Chats Work

When you chat in a workspace, you're not just chatting with a generic AI. Instead, you're:

• Searching through your specific document collection
• Retrieving relevant information
• Getting contextual answers
• Exploring your knowledge base interactively

Key Interactions

1. Document Search

   • Ask questions about your documents
   • Find specific information across multiple files
   • Get precise, source-backed answers

2. Contextual Understanding

   • The AI understands the context of your documents
   • Provides answers that reference specific sources
   • Can pull information from multiple documents

3. Knowledge Exploration

   • Summarize complex documents
   • Extract key insights
   • Compare information across different sources

Language Models (LLMs): The Brain of Workspace Chats

Language Models are critical because they:

• Interpret your questions
• Understand document context
• Generate human-like responses
• Interact with your document knowledge base

Adding a Language Model to Your Workspace

1. Navigate to Workspace Chats
   • Click on the "Chats" tab located at the top right of the workspace page

2. Access Model Selection
   • In the chat interface, look for "Default Model" in the top-left corner
   • Click on this to open model options

3. Add New Model
   • Select "Add Model" from the dropdown menu

4. Choose Your LLM
   • Available options:
     • OpenAI
     • Google
     • Anthropic

5. Provide API Credentials
   • Enter your API key for the selected model
   • API key is securely stored and managed within the workspace

Supported Language Models

OpenAI

• Models: GPT-3.5, GPT-4
• Best for: Detailed analysis, creative tasks

Google

• Models: Various Google AI models
• Best for: Comprehensive research, technical queries

Anthropic

• Models: Claude family
• Best for: Nuanced understanding, complex reasoning

Example Scenarios

Research Workspace

• "Summarize the key findings from these scientific papers"
• "Extract all mentions of climate change across these reports"
• "Compare different methodologies in these research documents"

Customer Support Workspace

• "Find all documents related to billing issues"
• "Summarize common customer complaints"
• "Retrieve specific support ticket resolutions"

Legal Compliance Workspace

• "List all regulatory changes mentioned in these documents"
• "Cross-reference compliance requirements"
• "Identify potential legal risks in our documentation"

Best Practices

1. Use clear, specific questions
2. Reference document names or types if possible
3. Utilize the model that best fits your workspace's domain
4. Regularly update and maintain your document collection

Limitations

• Answers are only as good as your uploaded documents
• Requires appropriate LLM and embedding model configuration
• Depends on document quality and comprehensiveness

---

Unlock the Power of Your Documents Transform your workspace into an intelligent, searchable knowledge base!