How to use the Data Hub in journeys

Overview

Insight mining is how you turn uploaded research into draft insights on a journey. You select the sources you want to use, choose where the insights should land, and Journey AI handles the extraction. This guide walks through how to set up and run a mining job from Data Hub.

Before you start

  • Your sources need to be fully Processed before they can be mined. Check the status column in your library — sources that are still processing or have errored can't be selected for mining.
  • You need insight edit permissions to run mining jobs and review results.
  • AI must be enabled for both your organization and your workspace.

Running a mining job

  1. Go to Data Hub in the left sidebar.
  2. Select the sources you want to mine. You can select up to 30 per run.
  3. Choose what to do with them:
    • Create journey – generate a new journey from scratch using the selected sources
    • Enrich journey – mine the sources into an existing journey

If you're enriching an existing journey

  1. Select the journey you want to enrich.
  2. TheyDo navigates you to that journey and opens the mining dialog.
  3. Configure your options:
    • Insight lanes – which lanes to populate with mined insights
    • Steps – optionally focus mining on specific steps
    • Insight intent – create new insights, or enrich ones that already exist
    • Granularity – how fine-grained the extracted insights should be
  4. Click Start mining.

Reviewing results

Mining runs in the background — Journey AI processes your sources asynchronously. When the job is done, a review flow opens where you can:

  • Accept insights to add them to your journey map
  • Reject insights you don't want to keep
  • Approve or reject in bulk

Accepted insights become visible on the journey. Rejected ones are discarded. Results that aren't reviewed stay in a pending state until you action them.

Tips

  • Only one mining job can run per journey at a time. If a job is already active, wait for it to finish before starting a new one.
  • Sources in an error or processing state can't be mined. Fix or re-upload them first, then try again.
  • You can set up recurring mining automation for connected sources (like Qualtrics or Medallia) so new data gets mined automatically as it comes in. Each journey supports up to five automations.
  • Pending results that go unreviewed for a long time will eventually expire. Try to review them promptly so nothing gets lost.