How to Analyze Survey Responses with AI

Sentiment analysis, automated tags, and automated summaries of responses allow your team to save time analyzing responses and get valuable survey insights at a glance. These features use AI to automatically categorize survey responses.

In this article, we'll look at:

• How to use sentiment analysis
• How to use automated tags
• How to create an automated survey responses summary
• Use cases
• FAQ

---

How to use sentiment analysis

The sentiment score is represented as positive, neutral, or negative. You'll also be able to see a breakdown of sentiment data for specific survey questions and see how the sentiment changes over time.

Enable the feature

Open your survey Responses.

Make sure that the survey has at least one short/long text question. The sentiment analysis does not work with other question types.

Switch the Sentiment toggle on.

View the sentiment scores.

Once enabled, your survey responses will be quickly analyzed and you'll see sentiment scores applied to each answer.

Edit sentiment scores

To edit a sentiment score, hover over a survey response and click the tag icon. You can then edit both the sentiment tags and the manual tags. To remove a sentiment score, hover over the sentiment score and click the small 'X' icon that appears within it.

Sentiment filters

You can filter your survey results based on sentiment by using the Sentiment filter. Learn more about filters.

Sentiment breakdown

To view the Sentiment breakdown page, click on the Results tab and select a specific short/long text question to see the sentiment data associated with it. Click the "Refresh data" icon in the upper right of the breakdown section to load the most recent results.

Back to top

How to use automated tags

Automated tags helps you detect recurring themes in your survey responses. You can define the tags yourself and let AI find the themes in your responses automatically.

Create tags

Open your survey Responses.

Make sure that the survey has at least one short/long text question. Automated tags don't work with other question types.

Switch the Automated tags toggle on.

You'll find it above your survey responses.

Enter your own tags or choose from the list.

Click Confirm tag selection.

Once you create your tags, responses will be tagged automatically.

Edit or delete tags

To edit or delete a tag, click Edit next to the Automated tags toggle.

Any changes you make will apply only to new responses. The tags assigned to existing responses will stay the same. You can delete tags manually for each response.

If Automated tags is toggled off, the tags created by it will be hidden. When re-enabled, the tags will reappear.

Filter by tag

You can filter your survey results based on tags to quickly find the most relevant responses. Learn more about filters.

Back to to

Tags breakdown

To view the Tags breakdown page, click on the Results tab and select a specific short/long text question to see the sentiment data associated with it.

Back to top

How to create an automated survey summary

To create a summary, your survey needs to have used long or short text question types and received at least 20 responses.

Open your survey.

Click Generate report.

Text responses from the last 30 days are used to generate the report.

Review your AI-generated report.

The report is divided into four sections:

• Summary: High-level summary of survey responses.
• Findings: Bullet points that describe and summarize common themes in submitted responses.
• Quotes: Relevant responses that enrich the findings with the user's voice.
• Next steps: Recommended actions based on the findings.

Once you generate a new summary, it'll be saved and available for all your team members. To access the report, open the survey and click Report > View report.

If you've gathered new responses and you want to edit or update an existing report, you can create a new updated version. Open the survey and click Report > Generate new report. A new automated report will be created based on text responses from the last 30 days. The previous version of the report won't be available after the update.

Back to top

Use cases

Here are a few examples of use cases that would benefit from the sentiment analysis and automated tags feature:

• A product team using surveys to get new feature ideas wants to quantify user requests to get buy-in from their team.

  Type the feature requests you want to measure the impact of and let automated tags show you how many customers request it. Turn on the sentiment analysis to see if these requests point to dissatisfaction or if customers are already happy with your product.

• A research team wants to scale up their research efforts.

  Turn on sentiment analysis and automated tags toggles to let AI do the manual work of categorizing responses based on user sentiment and tagging responses to save time on survey analysis. Use the charts in the Results tab directly in your presentations.

• A marketing team wants to understand if users are happy with the new homepage variant they’re testing and find out what to improve.

  Set up an on-site survey on the variant and control pages. Then turn on sentiment analysis to see the sentiment breakdown of both pages and compare the percentage of negative and positive answers. Use automated tags to quantify improvement ideas and understand what to work on in the next iteration.

• A UI/UX team is testing design alternatives and wants to understand which one their users like best and how they can improve it.

  Send two concept tests with different designs to the same audience. Then compare user sentiment for both to find out which your users would prefer. Use automated tags to enter some improvement ideas and see how many users mention them in their responses.

• A product team using the button surveys wants to detect bugs and feature suggestions quickly.

  Turn on the automated tags and enter some tag suggestions like “bug,” “UX issue,” or “improvement.” The AI will automatically tag your open-ended questions and show you how many responses have those tags so you can understand the impact of each request and prioritize brilliantly. Use sentiment analysis to measure the overall user sentiment on your product and see how it changes over time. Jump to the negative responses easily to pinpoint issues quickly.

Back to top

Back to top