#6 Office Hours
Hyper-Personalized Leads Engagement, Telegram Bot How-To & MCP 2.0 🚀
1. Introduction 0:00
Welcome to the session with 11 participants.
Purpose: To share updates and improvements in various topics.
2. Agenda Overview 0:19
Topics to be covered:
Knowledge engine revamp with a new embedding model.
Introduction of enterprise features to Azure Digflow.
Updates on MCP connection improvements.
Demo of hyper-personalized lead engagement.
Telegram bot tutorial.
3. Knowledge Engine Revamp 2:05
Transition from an older embedding model (rank 166) to a new one (rank 5).
New model features:
In-house development for better performance.
Increased embedding dimension from 768 to 4000.
Action required: Change relevance score from 0.5 to 0.2 for better results.
4. Enterprise Features in Azure Digflow 5:51
Introduction of re-ranking for retrieved results.
Explanation of hybrid strategies in agent settings:
Options for hybrid short, semantics, and full text searches.
5. MCP Connection Updates 7:53
New UI for MCP client connections.
Demo of creating connections with improved header management.
6. MCP Going Mainstream 10:26
ChatGPT now supports MCP, indicating a shift towards broader adoption.
More MCP providers expected to emerge.
7. Hyper-Personalized Lead Engagement Demo 15:34
Workflow for creating personalized invitations using LinkedIn profiles.
Demonstration of fetching LinkedIn data and generating emails.
8. Telegram Bot Tutorial 27:49
Step-by-step guide to creating a Telegram bot:
Setting up triggers and connecting to the bot.
Demonstration of sending messages and handling responses.
9. Q&A Session 01:00:00
Addressing questions from participants regarding features and functionalities.
10. Conclusion 01:09:30
Summary of the session and encouragement to reach out with further questions.
Link to Loom
Last updated
Was this helpful?