Skip to main content
Gainsight Inc.

Takeaways FAQs

This article addresses frequently asked questions about Takeaways.

This article contains common questions and answers about Takeaways.

How do I enable the Highlights in Takeaways for the first time?

Admins can enable Highlights in Takeaways. The initial setup may take up-to 72 hrs, after which an email notification is sent to the admin. For more information on how to enable Highlight in Takeaways, refer to the Overview of Horizon AI Toggle: Empowering User Data Control article.

What technologies is Gainsight using to generate Highlights? How is  data privacy being ensured?

Gainsight uses IBM Services and OpenAI ChatGPT API, through Microsoft Azure. Gainsight has enterprise agreements with these parties, ensuring that  all data remains in our instances and is not utilized for AI training purposes. For more information, refer to the Data Privacy FAQs for Generative AI Supported Features FAQs.

How are Key Opinions generated?

Gainsight uses IBM key point analysis service to analyze data to extract frequently expressed similar opinions from customers from Survey and/or Timeline source for a specific timeframe to derive key opinions.

How are categories assigned to key opinions?

ChatGPT API in MS Azure processes the key opinions to get categories.

How is the Overall Summary generated?

Overall Summary of What Works Well and What Needs Improvement is generated using ChatGPT through MS Azure on key opinions.

How are Key Opinion Summaries generated?

Key Opinion summaries are generated using ChatGPT through MS Azure on all Survey and Timeline data.

Can Highlights be shared with another user directly?

Currently, Gainsight has not created an option to share it.

How often is the Highlights data refreshed?

Highlights data undergoes a refresh every four hours, based on new data available from the Survey or Timeline. The last refreshed information is displayed in the top-right corner in the Text Analytics page.

Can users mark certain terms (for example, name of a product) as exempted from processing as positive or negative?

Currently, users don’t have a way to exempt specific words from sentiment analysis. However, our service is likely to interpret them within context.

  • Was this article helpful?