HTTP and Slack: AI Survey-Response Theme Extractor Template
A scheduled Spojit workflow that pulls raw survey responses from your survey tool's REST API, has an Agent-mode Connector node extract the recurring themes as structured output, and posts a clean summary to Slack.
What It Builds
This template builds a Spojit workflow on a Schedule trigger. On the cadence you set, a Connector node on the http connector calls your survey tool's REST API with http-get to fetch recent responses. An Agent-mode Connector node then reads the free-text answers, groups them into recurring themes, and returns a structured list. Finally a Connector node on the slack connector uses send-message to post the themes to a channel so your team sees what people are saying without anyone reading every response by hand.
The Prompt
Paste this into Miraxa and it builds the workflow, connecting the tools for you:
Build a workflow that runs every weekday at 9am. Use an http connector node to GET recent survey responses from my survey tool's REST API at https://api.mysurveytool.com/v1/responses, passing my API key in the Authorization header. Then use an Agent-mode Connector node to read the free-text answers and extract the top recurring themes, returning each theme with a short label, a one-line summary, and how many responses mentioned it as structured JSON. Finally post the themes to my #voice-of-customer Slack channel as a readable summary.
Connectors Used
- Schedule trigger - runs the workflow on a 5-field cron schedule with an IANA timezone (for example
0 9 * * 1-5). - http - calls your survey tool's REST API with
http-getto pull responses (works for any survey platform, no native tile needed). - Agent-mode Connector node - reads the answers and extracts recurring themes, with a Response Schema forcing structured JSON output.
- slack - posts the theme summary to a channel with
send-message.
Customize It
Change the cron expression in the prompt to shift the cadence (for example a Monday weekly digest), swap #voice-of-customer for your own channel, and point the API URL at your survey tool's real responses endpoint. You can also tune what the agent returns: ask it to include example quotes per theme, flag negative sentiment, or only surface themes mentioned by more than a set number of respondents. After Miraxa scaffolds the workflow, open each node's properties panel to confirm the Authorization header value and the Response Schema fields.
Tips
- If your survey tool's responses are paginated, ask Miraxa to loop through pages with a Loop node before the agent step so no responses are missed.
- Keep the survey API call in Direct mode (the
http-gettool is deterministic and costs no AI credits); reserve Agent mode for the theme extraction, where judgment is needed. - Define a tight Response Schema on the agent node (label, summary, count) so the Slack message formats cleanly every run instead of free-form prose.
Common Pitfalls
- Schedule triggers use the timezone you set, so a 9am cron in the wrong IANA zone posts at the wrong local hour. Confirm the timezone when Miraxa builds the trigger.
- The Slack channel must exist and your Slack connection must be able to post to it; for private channels, invite the connection first or
send-messagewill fail. - If the survey API returns no new responses, the agent has nothing to summarize. Add a Condition node to skip the Slack post when the response list is empty.
Related
- How to Analyze Customer Feedback at Scale with AI walks the full step-by-step build if you want to assemble this without the prompt.
- How to Connect to Any REST API Using HTTP Requests covers calling your survey tool's API with the http connector in detail.
- Using Connector Nodes in Agent Mode explains Response Schema and structured output for the theme-extraction step.