Skip to main content

Census Integration

Written by Team Enterpret

If you're using Census for syncing Users and Accounts, please follow the instructions on this page: https://helpcenter.enterpret.com/en/articles/8611269-syncing-users-and-accounts

You can use Enterpret's webhook APIs via Census/Fivetran Activations to ingest review, survey response, and conversation feedback data from various custom feedback sources into Enterpret. This integration allows you to quickly and seamlessly bring in feedback data from your warehouse or supported activation sources so that you can gain a more comprehensive understanding of customer feedback across all your channels.

In this article, we cover a sample use case of how Census/Fivetran Activations can be leveraged to ingest data from different sources into Enterpret.

To achieve this integration, we will first set up an Enterpret destination in Fivetran. We will then add a source and create an activation sync, which will automatically ingest the data into Enterpret.

Creating Destination

Open Fivetran and go to Activations.

  1. Go to Activation Destinations.

    Activation Destinations in Fivetran

  2. Add or open the Enterpret destination.

    Enterpret destination setup

  3. Provide a descriptive Name for your destination.

  4. Add your API Token.

    Add the Enterpret API token

  5. Run the destination connectivity test and confirm the destination is Healthy.

    Run the destination connectivity test

    Healthy Enterpret destination

If the Enterpret destination is not available in your Fivetran workspace, please reach out to your Enterpret team or Fivetran admin to enable the Enterpret partner destination.

To generate API Token

  1. Navigate to the Integrations page in your Enterpret instance.

    Integrations page in Enterpret

  2. Click on +New Integration, and select Webhook from the list by clicking its "Continue" button, then select "Feedback Integrations" as the integration category.

    Create a Webhook feedback integration

  3. Provide a Display Name for the integration in the presented form. This name will appear as the Source for records ingested using this Webhook.

    Add a display name for the Webhook integration

  4. Click Integrate, then Confirm and Integrate. A Webhook integration will be created with an API Key that you can copy.

    Copy the generated API key

Creating a Source

Go to Activation Sources and add the source that contains the feedback rows.

If you're using Google Sheets as the source, Fivetran creates a Google service account for the activation source. Share the spreadsheet with the generated service account and grant Editor access. Without this access, Fivetran source testing fails before it can read records.

Creating a Sync

  1. Go to Activation Syncs and create a new activation sync.

    Activation Syncs in Fivetran

  2. Choose your source connection, spreadsheet/table, and sheet/table name.

  3. Select the Enterpret destination.

  4. Select the object type from the drop-down menu.

    • Review Records: To ingest review feedback.

    • Survey Response Records: To ingest survey response feedback.

    • Conversation Records: To ingest conversation feedback.

    Select Conversation Records as the destination object

  5. For an initial test, use Create Only unless your use case requires a different behavior.

  6. Select id as the source sync key and Id as the destination sync key.

  7. Set up Custom Destination API Field Mappings.

To ingest Review Data

Set the source column to the Text field if it contains review text.

To ingest Survey Data

  1. Click on Add Mapping.

  2. Select the source column from the drop-down menu on the left.

  3. Provide the survey question as the field name with a feedback_ prefix.

To ingest Conversation Data

Use Conversation Records when each source row represents one conversation.

Map the required fields as shown below:

Source column

Destination field

Field API name

Type

Required

id

Id

id

string

yes

createdAt

Created At

createdAt

integer

yes

fileID

Source ID

fileID

string

yes

conversationMessages

Conversation Messages

conversationMessages

string

yes

Required Conversation Records field mappings

Conversation Messages must be a JSON string containing an array of messages. Each message must include an actor and text. Supported actors are user, agent, and bot.

Example:

[
  {"actor":"user","text":"Hi, I need help updating my profile."},
  {"actor":"agent","text":"Sure, I can help with that."}
]

Validation rules:

  1. conversationMessages must be a string.

  2. The string must parse as a JSON array.

  3. The array must contain at least one message.

  4. Each message must have actor and text.

  5. actor is trimmed/lowercased and must be one of user, agent, or bot.

  6. text is trimmed and must not be empty.

Rows where conversationMessages is not valid JSON, has an unsupported actor, or contains an empty message will fail validation and will not be ingested into Enterpret.

To ingest any additional fields along with the feedback

You can also ingest additional source columns along with the feedback as metadata.

  1. Click on Add Mapping.

  2. Select the desired source column from the drop-down menu on the left.

  3. Provide a field name for the metadata on the right.

  4. Choose the data type for the field. We currently support string and float types.

Metadata fields are optional. They do not replace the required fields for the selected feedback object.

Running a sync

Once you have defined the mapping, run a test sync and confirm the destination record succeeds.

Completed test sync with a successful destination record

Once you've got your sync up and running, you can configure your sync to run manually, on a schedule, or through the trigger mode that best fits your Fivetran setup.

For detailed instructions on how to create a periodic trigger, please refer to the Fivetran/Census docs.

Note: After the sync is finished, it can take up to 4 hours for the feedback to show up on Enterpret. A completed Fivetran sync confirms that records were sent successfully; downstream processing in Enterpret can take additional time before the feedback is visible in the UI.

Troubleshooting

Issue

What this usually means

How to fix it

Enterpret destination is not available in Fivetran

The Enterpret partner destination may not be enabled in your Fivetran workspace.

Reach out to your Enterpret team or Fivetran admin to enable the Enterpret destination for your workspace.

Destination test fails

The API token may be missing, invalid, or from the wrong Enterpret org.

Generate a new API token from an Enterpret Webhook Feedback Integration and paste it into the Enterpret destination setup. Then run the destination test again.

Conversation Records is not visible

Destination objects may need to be refreshed, or Conversation Records may not be enabled for the org tied to the API token.

Refresh destination objects in Fivetran. Confirm the API token belongs to the Enterpret org where Conversation Records is enabled.

Google Sheets source test fails

Fivetran cannot read the spreadsheet.

Share the spreadsheet with the exact Fivetran-generated service account and grant Editor access. Then test the source again.

Source columns are missing in the mapping step

Fivetran may not have refreshed the source schema yet, or the spreadsheet/table headers are not readable.

Confirm the source has column headers, refresh source fields, and re-open the mapping step.

Destination fields are missing in the mapping step

Fivetran may be showing stale destination object fields.

Refresh destination fields or destination objects, then return to the mapping step.

Records are rejected

One or more required fields are missing or conversationMessages is not valid.

Confirm id, createdAt, fileID, and conversationMessages are mapped. Check that conversationMessages is a valid JSON string, contains at least one message, uses only user, agent, or bot as actor values, and has non-empty text for every message.

Sync succeeds but feedback is not visible in Enterpret yet

Fivetran sent the records successfully, but Enterpret downstream processing may still be running.

Wait up to 4 hours for the feedback to appear in Enterpret. If it still does not appear, share the sync ID, source record ID, and destination result with the Enterpret team.

Metadata does not appear as expected

The additional source columns may not have been mapped as metadata fields, or the field type may not match the data.

Add each metadata field as its own mapping, choose the correct destination field name, and use string or float as the field type.

Give it a try and explore other use cases where you can leverage Enterpret's Census/Fivetran activation integration to ingest your customer feedback. We hope you found this tutorial helpful, and we look forward to seeing what you come up with.

Did this answer your question?