Comment Monitoring

Learn how to implement AI chatbots that can help users monitor, analyze, and respond to comments across their social media platforms using the MCP server.

Comment Monitoring & Response

This example demonstrates a simple workflow to monitor and respond to comments on social media platforms using the MCP server.

Use Case 1: Get Comments on Content

This use case focuses on retrieving all comments for a specific piece of content.

Getting Comments on Content

User: List all my users

AI: [Calls get-users-by-tenant-app]
Here are all your users:

1. John Smith ([e-mail])
   - ID: user-123-abc-456-def

User: List all user accounts John Smith has

AI: [Calls get-user-accounts with user_id]
John Smith has the following user accounts:

1. Instagram
   - ID: acct-456-insta-789
   - Status: Active

User: Get John's latest Instagram posts

AI: [Calls get-contents-by-user-account-id with user_account_id]
Here are John's 3 most recent Instagram posts:

1. Product Launch Announcement (May 5, 2025)
   - Content ID: content-123-456
   - URL: https://instagram.com/p/example1/

2. Behind the Scenes (May 1, 2025)
   - Content ID: content-789-012

User: Get comments on the Product Launch post

AI: [Calls get-comments with content_id]
Here are the comments on the Product Launch post:

1. @user123 (May 5, 2025): "When will this be available in Europe? Looks amazing!"
   - Comment ID: comment-123-456
   - Likes: 5
   - Replies: 0

2. @user456 (May 5, 2025): "What's the price point for this?"
   - Comment ID: comment-789-012
   - Likes: 3
   - Replies: 0

Implementation Steps

1. Get Users: Use get-users-by-tenant-app to retrieve all users
2. Get User Accounts: Use get-user-accounts to find the user’s social accounts
3. Get Content: Use get-contents-by-user-account-id to retrieve recent posts
4. Get Comments: Use get-comments with the content ID to retrieve all comments

Example Workflow

1. Identify the specific content you want to check comments for
2. Retrieve all comments for that content using its content ID
3. Display comments chronologically with relevant metadata (author, date, likes)
4. Show reply counts to identify comment threads

Use Case 2: Get Comments Based on Nature

This use case focuses on categorizing and filtering comments based on their content or intent.

Categorizing Comments by Nature

User: Get all comments on my Product Launch post and categorize them

AI: [Calls get-comments with content_id]
I've categorized the comments on your Product Launch post:

Questions (3):
1. @user123: "When will this be available in Europe? Looks amazing!"
2. @user456: "What's the price point for this?"
3. @user789: "Will there be a discount for existing customers?"

Praise (2):
1. @user234: "This product looks incredible! Can't wait to try it!"
2. @user567: "Your team always makes the best products. This looks amazing!"

Suggestions (1):
1. @user890: "You should add a blue color option too."

Complaints (0):
No complaints found in the comments.

User: Show me just the questions that need answers

AI: Here are the questions that haven't been answered yet:

1. @user123: "When will this be available in Europe? Looks amazing?"
2. @user456: "What's the price point for this?"
3. @user789: "Will there be a discount for existing customers?"

Would you like to respond to any of these questions?

Implementation Steps

1. Get Comments: Use get-comments or get-recent-comments to retrieve comments
2. Analyze Comment Text: Process the comment text to determine its nature:
   • Questions (contains ? or asking words like “how”, “what”, “when”)
   • Complaints (contains negative sentiment or problem descriptions)
   • Praise (contains positive sentiment or complimentary language)
   • Suggestions (contains words like “should”, “could”, “maybe”)

Example Workflow

1. Retrieve comments for the content
2. Categorize each comment based on text analysis
3. Group comments by category
4. Prioritize urgent categories (e.g., complaints or questions)

Use Case 3: Reply to Comments with Particular Emotions

This use case focuses on responding appropriately to comments based on their emotional tone.

Responding Based on Emotion

User: Show me comments with positive sentiment on my Product Launch post

AI: [Calls get-comments with content_id and analyzes sentiment]
Here are the positive comments on your Product Launch post:

1. @user234: "This product looks incredible! Can't wait to try it!"
   - Sentiment: Very Positive
   - Posted: May 5, 2025

2. @user567: "Your team always makes the best products. This looks amazing!"
   - Sentiment: Positive
   - Posted: May 6, 2025

User: Reply to the first positive comment

AI: [Drafts response based on positive sentiment]
Here's a suggested reply to @user234's positive comment:

"Thank you so much for your enthusiasm! We're excited too and can't wait for you to experience it. Stay tuned for the release date announcement coming next week!"

Would you like me to post this reply?

User: Yes, post it

AI: [Calls publish-comment with parent_id]
Reply successfully posted to @user234's comment:

"Thank you so much for your enthusiasm! We're excited too and can't wait for you to experience it. Stay tuned for the release date announcement coming next week!"

Implementation Steps

1. Get Comments: Use get-comments to retrieve comments for specific content
2. Analyze Emotion: Determine the emotional tone of each comment
3. Draft Responses: Create appropriate responses based on the emotional context
4. Post Replies: Use publish-comment with the parent comment ID to respond

Example Response Strategies

• Positive Comments: Thank the user and reinforce their positive experience
• Negative Comments: Acknowledge concerns, apologize if appropriate, offer solutions
• Neutral Questions: Provide clear, informative answers
• Excited Comments: Match their enthusiasm and build on their excitement

Use Case 4: Analyze Comment Sentiment Distribution

This use case focuses on understanding the overall sentiment distribution in comments.

Analyzing Sentiment Distribution

User: Analyze the sentiment of comments across all my Instagram posts this month

AI: [Calls get-contents-by-user-account-id and get-comments for multiple posts]
Here's the sentiment analysis for comments on your Instagram posts from May 1-7, 2025:

Total Comments: 42

Sentiment Distribution:
- Positive: 24 comments (57%)
- Neutral: 15 comments (36%)
- Negative: 3 comments (7%)

Posts with Highest Positive Sentiment:
1. Product Launch Announcement: 85% positive
2. Team Celebration Photo: 72% positive

Posts with Any Negative Sentiment:
1. Product Feature Update: 2 negative comments
   - "The new interface is confusing compared to before."
   - "Why remove the feature I used most?"

User: What topics are mentioned most in positive comments?

AI: [Analyzes text content of positive comments]
Top topics in positive comments:
1. Product quality (mentioned in 45% of positive comments)
2. Design aesthetics (mentioned in 30% of positive comments)
3. Customer service (mentioned in 15% of positive comments)
4. Price value (mentioned in 10% of positive comments)

Implementation Steps

1. Get All Comments: Use get-recent-comments or get comments for multiple content items
2. Analyze Sentiment: Categorize each comment as positive, negative, or neutral
3. Generate Statistics: Calculate the distribution of sentiments
4. Identify Trends: Compare current sentiment distribution with historical data

Example Analysis

1. Calculate the percentage of positive, negative, and neutral comments
2. Identify content with unusually high negative or positive sentiment
3. Track sentiment changes over time
4. Correlate sentiment with content types or topics

Required MCP Tools

This implementation uses these tools:

• get-users-by-tenant-app - Lists all users
• get-user-accounts - Lists user’s social media accounts
• get-contents-by-user-account-id - Gets recent content posted by a user
• get-comments - Gets comments for specific content items
• get-recent-comments - Gets recent comments across all content
• publish-comment - Posts replies to comments

Advanced Features

To enhance comment management capabilities, consider implementing:

• Automated sentiment analysis - Automatically detect the emotional tone of comments
• Comment prioritization - Flag urgent or important comments for immediate attention
• Response templates - Develop standardized responses for common comment types
• Engagement metrics - Track response rates and times across platforms
• Keyword alerting - Alert for comments containing specific words or phrases