All Collections
Survey campaigns
General info about Campaigns
Automated topic & sentiment classification
Automated topic & sentiment classification

Automatically analyze your text feedback and categorize it into industry-specific topics and identify the feedback sentiment.

Alex Bitca avatar
Written by Alex Bitca
Updated over a week ago

You can opt to have your survey responses analyzed and classified automatically into topics and sentiment. The automated feedback classification feature is available per survey campaign, which means that you can choose to process responses only from a specific survey or surveys instead of all of them.

Survey responses are processed using our integration with OpenAI's GPT system.

IMPORTANT: Retently is not sending our clients' or survey respondents' personal data to OpenAI. Only the survey text comments get sent to be processed in GPT and we're using IDs to match them with the corresponding survey response back in Retently.

To activate the automated feedback classification option, from the Campaigns page, access a campaign's edit mode and go to the "Analytics" tab.

By default, this option will be switched OFF. You can activate the automated topic and sentiment analysis by switching the toggle ON and choosing a relevant classifier for your industry.

NOTE: Currently, we support the following classifiers:

  • B2B Software & SaaS (English)

  • B2B Education & Training (English)

  • B2C Ecommerce (English)

  • B2C Financial Services (English)

  • Other (English) - which is a general classifier.

English is the main supported text feedback language. However, due to the integration with OpenAI, we also support other languages with varying levels of accuracy depending on GPT's training data.

Please keep in mind that currently, only the "Main open-ended" text comment is automatically processed and classified.

Once the Analytics feature is activated and the campaign is saved, any new text feedback received in this campaign will be analyzed automatically by our AI-powered algorithm, which will identify one or more topics and will apply them to the survey response, as well as identify the overall sentiment of the feedback.

A processed survey response with text feedback will include both topics and sentiment, as in this example:

Did this answer your question?