AI Client Chat

The system uses persistent assistants and conversation threads to maintain context. Each client has a dedicated AI assistant that remembers all conversations across all chat sessions. Each individual chat maintains its own thread ID that preserves conversation history, allowing the AI to reference previous messages and maintain context throughout the conversation. This means the assistant remembers not just messages within a single chat, but can recall information from previous chats with the same client.

Yes, you can create multiple chat sessions per client. Each chat maintains its own conversation history and context. However, the AI assistant is shared across all chats for the same client, meaning it can reference information from other chat sessions when relevant. This is useful for organizing different topics or discussions with the same client while maintaining overall context.

The document status will be set to "error" and you'll see an error indicator. You can try uploading the file again, or check if the file format is correct and not corrupted. Ensure the file is in a supported format and under 100MB.

No, the AI only has access to documents uploaded for the specific client you're chatting with. Client isolation ensures that documents from one client are never accessible when chatting with another client.

Yes, you can delete documents at any time. When deleted, the file is removed from both local storage (or S3) and OpenAI's servers, ensuring complete removal. The deletion is logged in the audit trail.

Processing time varies depending on file size and OpenAI's processing queue. The system automatically checks the status and updates when processing is complete. Typically, small files process in seconds, while larger files may take a few minutes.

Yes, if team sharing is enabled. Clients can be shared with team members, allowing multiple users to chat with the same client and access the same documents. Team sharing is optional and controlled by team settings.

Be specific and provide context. Good questions: "What is the interest rate mentioned in the loan agreement for John's mortgage?" Poor questions: "What's the rate?" The more specific and contextual your questions, the better the AI can search through documents and provide accurate answers. Since the assistant remembers previous conversations, you can also reference earlier discussions: "Based on our previous conversation about the mortgage terms, what was the final interest rate?"

Yes! You can upload documents directly without creating a client first. The system will automatically scan each document for client names, create new clients when names are found, extract relevant information, and save it. This is especially useful when processing multiple documents at once. Each file is processed individually: upload → scan/extract → create client (if name found) → save info → repeat for next file.

Yes! Each client has a persistent AI assistant that remembers all conversations across all chat sessions. This means if you ask about something in one chat session, the assistant can recall that information in future sessions. The assistant maintains context and builds on previous interactions, making conversations more natural and efficient over time.

Yes, you can switch between OpenAI and Gemini AI providers in the admin settings. However, each client's assistant is provider-specific. If you switch providers, new conversations will use the new provider, but existing assistants and their conversation history remain with their original provider. This ensures conversation continuity and prevents data loss.