Use natural language to create targeted segments


This content is archived and is not being updated. For the latest documentation, go to What's new in Dynamics 365 Marketing. For the latest release plans, go to Dynamics 365 and Microsoft Power Platform release plans.

Enabled for Public preview Early access General availability
Admins, makers, marketers, or analysts, automatically Oct 1, 2020 - Oct 1, 2021

Business value

You can now use natural language sentences to build segments faster in Dynamics 365 Marketing. Use simple words to specify the attributes and logic for the segment you want to target so you can reach the exact audience you are looking for. For example, using natural language, you can create a segment of “Contacts who were born between 1970 and 2000 and make more than $50K”.

Feature details

Natural language queries enable the creation of Marketing segment query blocks using natural language, simplifying the creation of complex segment queries. Using natural language makes it faster and easier to get started creating a segment.

Natural language input enables you to use common words and phrases to describe the people you'd like to include in a segment. Once you type and submit a statement, the service will fill the query fields using information collected from your contacts, accounts, and leads.

Here are examples of how to describe segments so you get the best results:

  • "Contacts living in Seattle or Chicago"
  • "Contacts with a birthday before next month"
  • "Contacts who were created last month"
  • "Contacts who prefer to get emails"


  • Use multiple query blocks. It's best if each query block has its own short query. You can add multiple query blocks, each with its own natural language query to build more complex segments.
  • Give feedback. By giving a "thumbs up" or "thumbs down" when your natural language query works (or doesn't work) as intended, you help to improve the AI. Your feedback and suggestions play a big part in improving the answers you get. Your suggestions help train the search to accurately identify context to better predict answers. If you get an answer that isn't what you expected, providing a suggestion will directly impact the model's ability to provide better, more relevant results in the future.


  • Natural language input works with Microsoft Dataverse using common entities. Custom entities are not currently supported.
  • Natural language processing and machine learning are dynamic, meaning they learn and change over time. The processing is able to identify phrases and expressions that may change from role to role, company to company, and industry to industry.

This feature is only available in outbound marketing; it does not affect real-time marketing functionality.

Natural language segment Contacts who were born between 1970 and 2000 and make more than $50K.

Geographic areas

This feature will be released into the following Microsoft Azure geographic areas:

  • United States
  • Europe
  • United Kingdom
  • Australia
  • Canada

See also

Use natural language in real-time marketing journeys (docs)