Interacting with the Research Chat

Interacting with the Research Chat

Synthesis provides an intelligent Research Chat feature, allowing you to have a dynamic conversation with your uploaded documents and agent-generated insights. Think of it as your personal AI research assistant, ready to answer questions, summarize information, and help you navigate the complexities of your project data.

The Research Chat is designed to accelerate your understanding and streamline your analysis by providing instant, context-aware responses.

Accessing the Research Chat

To interact with the Research Chat:

1. Navigate to the Dashboard.
2. Select an existing project by clicking on its card. This will open the Project Details View.
3. Within the Project Details View, look for the Chat with Research tab. Click on this tab to open the chat interface.

How it Works

The Research Chat leverages a Retrieval Augmented Generation (RAG) approach:

1. Context Retrieval: When you ask a question, Synthesis first performs a semantic search across all documents uploaded to your project. It identifies the most relevant passages and also considers recent outputs from your project's AI agents (e.g., from the 'writer' or 'outliner' agents) to provide comprehensive context.
2. AI Generation: This retrieved context, along with your conversation history, is then sent to a powerful Language Model (powered by Google Gemini). The AI processes this information to generate a helpful, accurate, and relevant response to your query.
3. Source Attribution: Whenever possible, the AI will indicate which source (e.g., a specific document or an agent output) it used to formulate its answer, helping you verify information.
4. Conversation History: The chat maintains a short memory of your previous questions and the AI's responses, allowing for more natural, follow-up conversations.

Asking Questions

Engaging with the Research Chat is straightforward:

1. Type Your Query: In the chat input box at the bottom of the interface, type your question.
   • Example questions:
     • "Summarize the key findings from all documents."
     • "What methodologies were discussed in the paper by Smith et al.?"
     • "Can you elaborate on the concept of 'semantic clustering' mentioned in the literature review?"
     • "What are the main hypotheses generated for this project?"
     • "Compare the results presented in Document A with Document B."
2. Send Your Message: Press Enter or click the send button to submit your question.
3. Review the Response: The AI assistant will process your request and display its answer in the chat window. Look for source indicators if the AI references specific content.

Understanding Responses and Limitations

• Contextual Answers: The AI will strive to provide answers based only on the context it retrieves from your project's documents and agent outputs. If it cannot find relevant information, it will inform you.
• Source References: Pay attention to any mentions of [Source X] or [Agent Y] within the response. This indicates where the information was drawn from.
• Processing Status: If your documents are still being processed (e.g., PDF text extraction is underway), the chat might inform you that it doesn't have access to the document content yet. In such cases:
  • Ensure your documents have finished uploading and processing.
  • Wait a few moments and try again.
  • Verify that documents were uploaded successfully.

Tips for Effective Chat Interaction

• Be Specific: Clear and concise questions often yield the best results.
• Ask Follow-Up Questions: Leverage the chat's memory to dive deeper into a topic. If an answer isn't fully clear, ask for clarification.
• Iterate and Refine: If the initial response isn't what you expected, try rephrasing your question or breaking it down into smaller parts.
• Check Processing Status: If you're consistently getting responses indicating a lack of context, ensure your documents are fully processed within the project.