Feature Request: Ability to Extract Answers to Semantic Questions from Conversations #3
Description
Summary
Add functionality to extract answers from conversations based on semantic questions. This feature would enable the system to analyze conversation data and directly return answers to user-defined natural language queries.
Motivation
Users often need to extract specific information from long conversations without manually searching through them. A semantic question-answering ability would make this process faster and more accurate by leveraging conversational context and AI-driven extraction.
Expected Outcome
- Users can input a semantic (natural language) question.
- The tool will analyze a given conversation and return the answer(s) from the dialogue.
- Extraction leverages context and is robust to conversational ambiguity.
Example Usage
- Question: "What were the main blockers reported by the team last week?"
- System Response: Extracts and outputs summarized blockers mentioned in relevant parts of the conversation.
Additional Notes
- Should support a variety of semantic question types (who, what, when, why, how, etc).
- Should perform well on both short and long-form conversations.
- Consider integration with existing AI/NLP modules in the project.
- Vectorization of chat history