AI Response Summarisation
Use AI to summarise open-text responses across all submissions into key themes and insights.
Open in appAI Response Summarisation analyses all open-text responses for a given field and produces a structured summary of key themes, common sentiments, and notable outliers — without you having to read each response individually.
Available on: Pro and Business plans.
How to use it
- Open the Responses tab for your form
- Click the column header for a Long Text or Short Text field
- Click Summarise with AI
- The AI analyses all visible responses (respects your current filters) and returns a summary
What the summary includes
- Main themes — the 3–5 most common topics mentioned across responses
- Sentiment breakdown — positive / neutral / negative distribution
- Notable responses — unusually detailed or outlier responses worth reading
- Suggested actions — where the AI identifies clear patterns that suggest a follow-up action
Example
For a post-event feedback form with 200 responses to "What could we improve?", the AI summary might return:
Main themes:
- Venue temperature too warm (mentioned in ~38% of responses)
- Networking breaks too short (mentioned in ~24%)
- Request for more technical sessions (~19%)
Sentiment: 52% positive, 28% neutral, 20% negative
Suggested actions: Consider venue with better climate control; extend networking breaks by 15 minutes.
Privacy and data handling
When you request a summary, the text content of responses in scope is sent to the Anthropic Claude API for processing. This applies to non-encrypted responses only.
For Storage Mode (encrypted) forms, AI summarisation is not available — you would need to decrypt responses first, and even then the summary feature is disabled to avoid sending sensitive decrypted data to external APIs.
Filtering before summarising
Use the response table filters (date range, search) before running the summary to scope the analysis. For example, summarise only last month's responses, or only responses containing a specific keyword.
Re-running summaries
The summary is generated on demand and is not cached. Run it again after new responses come in to get an updated analysis.
Something unclear? Let us know.