You can connect Gong to Harvestr to automatically create product feedback when a specific AI Tracker is detected in recorded calls. This integration works by sending a webhook from Gong to Harvestr whenever a call matches your AI Tracker.
Prerequisites
Admin access to Gong
An AI Tracker called "Product feedback" in Gong
A Harvestr Elite subscription with an Editor account
Note: Gong renamed Smart Trackers to AI Trackers, and moved them into Agent Studio. If you set up this integration before the change, your existing tracker still works — it now appears as an AI Tracker.
Step 1: Create an AI Tracker called "Product feedback"
Harvestr only imports calls that contain valuable product feedback.
To do this, we rely on Gong's AI Trackers to detect when product feedback is mentioned and extract the relevant insights from the call transcript.
In Gong, go to the Admin center > Agent Studio, find AI Tracker and click Settings. From there, create a new AI Tracker called "Product feedback".
You can also follow Gong's documentation for the full setup reference.
To ensure Harvestr imports only high-quality product feedback from your Gong calls, we strongly recommend setting up your AI Tracker exactly as described below.
1. Define the concept you want to track
In the free text question field, enter the following question:
Did the customer provide feedback or request features to improve the product?
2. Add details to guide Gong's AI
Copy this into the Details section:
Product feedback refers to any part of a conversation where a customer provides specific input on the product's existing features or expresses needs for new functionality. This includes negative experiences with current features, requests for new capabilities, comparisons with competitors, or concrete suggestions for improvement.
However, do NOT extract as product feedback any statements where the conversation context (including responses from a teammate or the company representative) makes it clear that the question or comment is about already existing capabilities that the customer was simply unaware of. For example, if a customer asks if a capability exists and is promptly informed by a teammate that it does, do not treat this as product feedback about missing features or functionality.
Do not consider as product feedback:
• Statements that only describe or praise existing capabilities
• Generic positive comments with no actionable input
• Basic product usage questions
• Expressions of satisfaction without specific feedback
• Greetings or small talk
• Administrative or procedural questions
• Questions clearly answered in the transcript as being already possible with current features
Examples of valid product feedback:
1. We're finding the process for adding new team members a bit clunky. Having to go through three separate menus to grant permissions is inefficient for our admins.
2. It would be amazing if we could filter search results by date range or by a custom field.
3. Do you have plans to add a Slack integration?
Examples that should not be captured:
1. I have to say, the new reporting dashboard is a game changer for our weekly meetings.
2. Everything works very well.
3. How can I export dashboards?
3. Choose tracking parameters
Whose side to track: Customer
Conversation types: Calls
Call scope: External
Step 2: Connect Gong to Harvestr
In Harvestr, go to Settings > Integrations > Gong
Click Connect to Gong
A window will open asking for permission to authorize the integration
Click Allow
Close the window after authorization
Copy the webhook URL and open Gong.
✅ Gong is now connected. You can move on to configuring the automation rule in Gong.
Step 3: Create an Automation rule in Gong
In Gong, go to Company Settings
Scroll down to Automation Rules, click Add Rule
Configure the rule as follows:
Trigger: New call
Filter: add your Product feedback AI Tracker as a filter with the 'Mentioned by customer' option.
Action: Fire webhook
URL: Paste the webhook URL copied from Harvestr
Authentication method:
URL includes keyName: e.g., "Harvestr" (this can be anything)
Click Test now to confirm the webhook is working
If the test succeeds, click Save
✅ Your webhook is now active and ready to send call data to Harvestr.
"Test passed successfully" does not send a call to Harvestr
Gong's Test now button only checks that the webhook URL is reachable and correctly authenticated. It does not send a real call. Seeing "Test passed successfully" and then finding nothing in your Harvestr inbox is expected behaviour, not an error.
The integration is not retroactive either: it does not import your existing Gong calls. Harvestr receives a call once all of these are true:
• The call was recorded after you saved the automation rule
• The "Product feedback" AI Tracker was applied to it
• The tracker detected product feedback mentioned by the customer
To verify the setup end to end, record a new call in which a customer clearly gives product feedback, then check your Harvestr inbox once Gong has finished processing the call.
Step 4: What Harvestr imports from the call
When a call is sent to Harvestr via the integration, we automatically create a new feedback item that includes:
🎯 Highlights from the "Product feedback" AI Tracker
📝 Summary
✏️ Transcript
🤵💼 Call participants
📅 Call date and time
🔗 Link to the full call recording
This ensures you see exactly what was said, by whom, and in what context.
Only feedback from the Highlights section can be categorized and linked to Discoveries.
For each new call, Harvestr's AI automatically parses the Highlights to extract relevant feedback and categorize it, saving you time and surfacing what matters most.