Chat Interface

The chat interface is where you ask questions and get AI-generated answers from your documents. This guide explains how to use it effectively.

Opening Chat

There are two ways to access chat:

Collection Chat

1. Open a collection
2. Click the Chat button

This searches only within that collection's documents.

Global Chat

1. Click Chat in the main navigation

This searches across all collections you have access to.

Asking Questions

1. Type your question in the input field at the bottom
2. Press Enter or click the send button
3. Wait for the AI to generate a response (typically 2-5 seconds)

Example Questions

Factual questions:

• "What is our refund policy?"
• "How many vacation days do employees get?"
• "What are the system requirements?"

Explanatory questions:

• "Explain the onboarding process"
• "How does the authentication system work?"
• "Describe the approval workflow"

Comparative questions:

• "What's the difference between Plan A and Plan B?"
• "Compare the old policy to the new policy"

Summary requests:

• "Summarize the key points of the employee handbook"
• "Give me an overview of the Q4 report"

Understanding Responses

Each AI response includes:

The Answer

A natural language response based on your documents. The AI synthesizes information from relevant sections to answer your question.

Sources

Below the answer, you'll see the sources used:

• Document name - Which file the information came from
• Page number - The specific page (for PDFs)
• Relevance score - How closely the content matches your question (shown as a percentage)

Viewing Source Text

Click on any source to expand it and see the exact text passage that was used. This helps you verify the answer and find more context.

Conversation History

The chat maintains a history of your questions and answers within a session.

• Scroll up to see previous messages
• Follow-up questions can reference previous context (e.g., "Tell me more about that")
• Start fresh by clicking "New Chat" to clear history

Note: Conversation history is stored locally in your browser and is not saved permanently.

Filtering by Tags

If your documents have tags, you can filter which documents are searched:

1. Click the Filter button (funnel icon)
2. Select one or more tags
3. Ask your question

The AI will only search documents that have ALL selected tags.

Example use cases:

• Filter to "HR" tag for policy questions
• Filter to "2024" tag for current year documents
• Filter to "Engineering" + "Backend" for specific technical docs

What the AI Can and Can't Do

The AI Can:

• Answer questions based on your uploaded documents
• Synthesize information from multiple sources
• Quote specific passages with citations
• Admit when it doesn't have enough information

The AI Cannot:

• Answer questions about topics not in your documents
• Access external websites or databases
• Remember conversations after you close the browser
• Make up information (it's designed to only use your documents)

Response Time

Typical response times:

Factors that affect speed:

• Number of documents in the collection
• Complexity of the question
• Length of the response

Tips for Better Results

1. Be specific: "What is the return window for electronics?" is better than "Tell me about returns"

2. Ask one thing at a time: Break complex questions into parts

3. Use keywords: Include terms that likely appear in your documents

4. Check sources: Always verify answers by clicking through to the source text

5. Rephrase if needed: If you don't get a good answer, try asking differently

See Tips for Better Answers for more techniques.

Keyboard Shortcuts

Related Articles

• Understanding Citations
• Tips for Better Answers
• Common Issues

Need Help?

If the chat isn't working as expected, check our troubleshooting guide or contact support.