<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/97eb35e736cf449ea22ccd83af9168b9&quot; frameborder=&quot;0&quot; width=&quot;1450&quot; height=&quot;1087&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1087</height><width>1450</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1087</thumbnail_height><thumbnail_width>1450</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/97eb35e736cf449ea22ccd83af9168b9-2506651c6341d991.gif</thumbnail_url><duration>635.648</duration><title>Build a Personalized AI Automation Bot</title><description>This Loom walks through a personal Make.com project that builds a Telegram-based AI virtual assistant with routing for different user intents. It uses a Grok intent router to branch into image processing with Gemini 1.5 or Gemini Flash (currently on the free tier), and document automation that generates structured Google Docs from conversational prompts. In a live demo, the bot creates a detailed Loom recording guide in Google Docs and also parses sales deal messages using a custom JSON schema to fill a Google Sheet, tracking fields like lead name, deal value, and status such as Close1. The author also shows a location tracking feature and notes the bot can handle basic conversations, with version 1 already released.</description></oembed>