This workflow combines AI-powered document question answering (QA) with an automated appointment scheduling assistant. It enables users to query PDFs (such as whitepapers) using natural language, powered by vector embeddings and GPT models, and also allows seamless scheduling of meetings via Google Calendar.
It’s a showcase of how to integrate LangChain, OpenAI/Anthropic models, Pinecone vector databases, and Google Calendar APIs into a single automation pipeline.
Key Features
- 📄 PDF QA with RAG
- Download and load PDFs with metadata.
- Split content into chunks and create vector embeddings (OpenAI).
- Store and retrieve data via Pinecone for context-aware QA.
- Answer questions only using document knowledge (RAG).
- 🤖 AI-Powered Chat Interface
- Chat with documents through GPT-4o or Anthropic models.
- Interactive conversation memory ensures context retention.
- Initial chatbot greeting pre-configured for appointment flow.
- 📅 Appointment Scheduling Assistant
- Uses Google Calendar API for availability checks.
- Automatically books 30-minute meetings if slots are free.
- Handles user requests naturally (e.g., “Book a meeting tomorrow at 10 AM”).
- ✉️ Smart Email Categorization
- Gmail integration with AI classification (e.g., automation vs. music).
- Automatically applies labels to relevant emails.
- 💬 Notifications
- Slack integration to send real-time updates when emails or triggers match certain conditions.
