- TRENDS
- BATCHES
- REPORTS
- HUMAN
Short answer
How can I use my AI to track customer sentiment trends?
You can send bulk datasets to your AI for sentiment scoring. This lets you spot mood shifts across thousands of entries at once. You'll see the specific steps to set this up below.
The outcomes
What you get from sentiment analysis
Running these tasks turns raw feedback into clear mood data.
TRENDS
Identify mood shifts
The AI flags when sentiment moves from positive to negative. You'll see exactly when customer frustration starts to climb.
BATCHES
Process bulk data
Instead of checking one-by-one, you run large groups of reviews. This gives you a high-level view of your whole customer base.
REPORTS
Track job status
You can check on the progress of long-running sentiment jobs. This keeps you updated on when your data is ready for review.
HUMAN
Verify edge cases
The AI handles the heavy lifting, but you decide which specific outliers or nuanced complaints need a personal response.
The workflow
What your AI does when the data arrives.
The AI handles the heavy lifting of processing text so you can focus on the meaning.
Start the analysis
Your AI sends your collection of customer feedback to the connector to begin the scoring process.
create_batch_predictionMonitor progress
If you are processing thousands of rows, your AI checks the status of the job to see if it is finished.
get_batch_predictionReview all jobs
You can ask your AI to look back at previous sentiment runs to compare current moods against older data.
list_batch_predictionsExtract insights
Once the job is done, your AI reads the results to tell you exactly what customers are feeling.
get_batch_prediction
Try it
Copy these to start.
Use these prompts to kick off your first sentiment analysis.
Starting points
These are just ways to start the conversation. Swap out the bracketed text for your actual data sources.
Deep Talk Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_p1r9lyvKUygLDYkDHtS0iQrZChD9ZM45aBPc7pHA/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Deep Talk capabilities are ready to use.
{
"mcpServers": {
"deep-talk-mcp": {
"url": "https://edge.vinkius.com/vk_preview_p1r9lyvKUygLDYkDHtS0iQrZChD9ZM45aBPc7pHA/mcp"
}
}
}Run a sentiment prediction on the feedback in my 'Q3_Reviews' Google Sheet.
Check the status of my last sentiment batch prediction to see if it's done.
List my recent batch predictions and tell me if the customer mood is improving.
Analyze the sentiment for the latest batch of support tickets from my Zendesk export.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Deep Talk once, then ask.
Just click the link to connect. Your credentials stay encrypted and private, and you can immediately start talking to your AI about your data.
Connect Deep Talk to your AIFAQ
How this task behaves
- 01
Can the AI automatically reply to negative reviews?
No. The AI only predicts sentiment and analyzes trends. You must handle all customer communications yourself.
- 02
How many reviews can I analyze at once?
The AI uses batch predictions to handle large datasets, allowing you to process massive amounts of text in a single job.
- 03
Does the AI delete my original feedback data?
No. The AI only reads the data you provide to generate predictions. It cannot modify or delete your source files.
- 04
Can I see the progress of a large sentiment job?
Yes. You can ask your AI to check the status of a specific prediction job while it is running.
- 05
Will the AI change the sentiment scores once they are generated?
No. The scores are based on the data provided during the prediction run and cannot be edited by the AI.
More questions about Deep Talk? The Connector page answers them. See everything the Deep Talk Connector can do
Connect Deep Talk to Claude, Cursor, ChatGPT & more
