Monitor brand mentions on X with Gemini AI visual analysis & Telegram alerts

This workflow automates brand monitoring on X by analyzing both the text and the images in posts. It uses multi-modal AI to score brand relevance, filters out noise, logs important mentions in Airtable, and sends real-time alerts to a Telegram group for high-priority posts.

What it does

Traditional brand monitoring tools often miss the most authentic user content because they only track text. They can't "see" your logo in a photo or your product featured in a video without a direct keyword mention.

This workflow acts as an AI agent that overcomes this blind spot. It finds mentions of your brand on X and then uses Google Gemini's multi-modal capabilities to perform a comprehensive analysis of both the text and any attached images. This allows it to understand the full context of a mention, score its relevance to your brand, and take the appropriate action, creating a powerful "visual intelligence" system.

How it works

The workflow runs on a schedule to find, analyze, and triage brand mentions.

1. Get New Tweets: The workflow begins by using an Apify actor to scrape X for recent posts based on a defined set of search terms (e.g., Tesla OR $TSLA). It then filters these results to find unique mentions not already processed. 2. Check for Duplicates: It cross-references each found tweet with an Airtable base to ensure it hasn't been analyzed before, preventing duplicate work. 3. Analyze Post Content: For each new, unique post, the workflow performs two parallel analyses using Google Gemini: Analyze the Photos: The AI examines the images in the post to describe the scene, identify logos or products, and determine the visual mood. Analyze the Text: A separate AI call analyzes the text of the post to understand its context and sentiment. 4. Final Relevance Check: A "Head Strategist" AI node receives the outputs from both the visual and text analyses. It synthesizes this information to assign a final brand relevance score from 1 to 10. 5. Triage and Action: Based on this score, the workflow automatically triages the post: High Relevance (Score > 7): The post is logged in the Airtable base, and an instant, detailed alert is sent to a Telegram monitoring group. Medium Relevance (Score 4-7): The post is quietly logged in Airtable for later strategic review. Low Relevance (Score < 4): The post is ignored, effectively filtering out noise.

Setup Instructions