This week brings more precise matching to RealTime B2B Customers data products, targeted webhook filtering by data product, and a new way to retrieve table schemas programmatically - plus another look at the Duplicate Handling controls coming soon!

Filter Webhook Record and Curation Events by Data Product

You can now configure your webhooks to receive record and curation events from only selected data products! Receive events only from the data products you are interested in monitoring - reducing the number of events you receive and removing the need to filter these events downstream.

This feature is available for both existing and new webhooks.

See Configuring Webhooks for details.

Improved Matching in RealTime B2B Customers Data Products

We've made several improvements to increase match precision in RealTime B2B Customers Data Products.

• Tamr now extracts and compares legal form (Inc., LTD., and so on) in company name values. Tamr then compares the company name without legal form and the legal form itself when identifying matching records. As part of this change, Tamr returns these new match labels:
  • MATCH_LEGAL_FORM: The legal form extracted from the company name value was a match.
  • MATCH_CROSS_LEGAL_FORM: The legal form extracted from the company name value in one record and the alternative company name value in another record match.
  • NO_MATCH_LEGAL_FORM: The legal form extracted from the company name value was not a match.
  • MISSING_LEGAL_FORM: The legal form could not be extracted from the company name value.
• Tamr now identifies matches or partial matches for primary full addresses only for full addresses that include address line 1.
  
  As part of this change, Tamr returns a new MATCH_SPARSE_PRIMARY_ADDRESS match label. This label indicates a match on full address between two records in which the primary full address does not include address line 1.
• When comparing primary address line 1 values, Tamr now extracts the street number (if provided) and identifies whether the number matches or does not match between records. As part of this change, Tamr returns these new match labels:
  • MATCH_PRIMARY_HOUSE_NUMBER: The number extracted from primary address line 1 was a match.
  • NO_MATCH_PRIMARY_HOUSE_NUMBER: The number extracted from primary address line 1 was not a match.
  • MISSING_PRIMARY_HOUSE_NUMBER: The number could not be extracted from primary address line 1.

Important Note: These updates are applicable for new data products only; they are not included in data products created before October 1, 2026. If you are interested in updating an existing data product with these changes, contact Tamr Support ([email protected]).

Get a Table's Schema Programmatically

A new get schema endpoint is now available in the Tables API. Use this endpoint to retrieve a table's schema as a JSON Schema document conforming to the https://json-schema.org/draft/2020-12/schema dialect.

Coming Soon! Fully Control How Tamr Handles Possible Duplicates

In the next few weeks, we'll be releasing a new Duplicate Handling page in RealTime data product configuration.

This page will allow you to:

• Choose the default merge behavior for your data product and decide how to handle suggested duplicate records based on match confidence levels: merge them automatically, send them to a curation queue for review, or ignore them.
• Create multiple suggested duplicate queues to apply different duplicate handling rules to different kinds of records.
• Route suggested duplicates to specific curation queues based on attribute values (for example, send high priority accounts or records from a specific region to separate queues).
• Automatically dismiss stale queue items to help curators focus only on relevant items.

Bug Fixes and Other Improvements

• Added general usability, styling, and performance improvements.