Pull surveys and feedback from Databricks

Reviewer note: Beta feature (shipped Jul 2026), available on Strategic plans.

Overview

This article is for teams who already have Databricks connected to TheyDo and want to bring qualitative data, such as survey responses and open-text feedback, into the Data Hub. Until now the Databricks integration pulled numeric data for metrics. You can now also import text responses as Quotes, map them to your taxonomy, and let AI mine them into insights. The source refreshes daily, so new feedback keeps flowing in without manual re-imports. This article covers how to add a survey data source, map your fields, run the sync, and start working with the data.

Note: Importing surveys and feedback from Databricks is a beta feature, available on Strategic plans. If you would like access, contact your Customer Success Manager.

Before you start

You need your Databricks connection already set up in TheyDo. If you have not connected Databricks yet, follow Databricks integration first, then come back here. You also need the SQL warehouse you plan to use to be able to reach the table that holds your survey or feedback responses.

Tip: Work with your data team to confirm which table and columns hold the response text and the attributes you want to tag on, before you start. A quick check now saves a re-sync later.

Add a survey data source

1. In TheyDo, go to Settings > Integrations.
2. Select your Databricks instance and click Manage.
3. Add a new data source and give it a recognizable name, such as My Survey.
4. Select Survey as the data source type.
5. Choose the SQL warehouse to query and the destination workspace the data should land in.
6. Set the lookback period to control how much historical feedback is imported. For example, choose three months.
7. Enable Convert all rows to Quotes when each row in your results is a piece of feedback you want in the Data Hub.

Query and preview the data

1. Enter the SQL query that returns your survey or feedback records.
2. Run a preview.
3. Check the preview to confirm the response text and the metadata you expect are both present before you continue.

Tip: Return only the columns you actually need: the text you want mined, plus the attributes you want to tag on (such as region, segment, or persona). A focused query makes the mapping step in the next section much simpler.

Map your data to your taxonomy

Map each column your query returns to the matching part of your TheyDo taxonomy. Column names in Databricks do not need to match your taxonomy names, so you do not have to rename anything in your source table.

Typical mappings look like this:

Databricks column	Maps to in TheyDo
country	Region tag group
segment	Portfolio Segment tag group
NPSReason	Text to mine
ImprovementSuggestion	Text to mine
Persona	Persona tag group

Note: A single response can hold more than one piece of useful feedback. Map each relevant text column to Text to mine so every one of them becomes a Quote.

Start and verify the sync

1. Review the source configuration and your field mappings.
2. Start the sync.
3. Open the Data Hub and wait for the new source to appear and finish generating Quotes.
4. Open the source and confirm the Quotes and taxonomy tags were created as you expected.

Work with the imported feedback

Once the sync finishes, your survey responses live in the Data Hub as Quotes, ready to use the same way as any other Data Hub source. You can:

• Explore and analyze the Quotes with the TheyDo Agent.
• Select the source when you create a new journey.
• Use the source to enrich an existing journey.
• Run AI mining to turn the feedback into insights mapped onto journey steps.

The source refreshes daily, so new and updated responses from Databricks keep flowing into TheyDo, and AI can keep surfacing insights as fresh feedback arrives.

Tip: Bringing operational data and customer feedback from the same Databricks warehouse into TheyDo means you can sit the numbers (your metrics) next to the reasons (your feedback) on the same journey.