Introducing intelligence in relevance search

We are thrilled to share that relevance search is becoming more intelligent through the infusion of machine learning. Moving forward, relevance search results will improve over time, thus making it easier to find the data you’re looking for. Today we are sharing some new capabilities, which will continue to evolve with increasing usage:

Spelling check

If a search query is misspelled, relevance search will attempt to find close matches based on what was typed.

Recognizing common abbreviations and acronyms

Relevance search will recognize common abbreviations and acronyms in search queries. The abbreviations and acronyms supported will grow over time.

Better understanding

Relevance search will try to better understand the query by applying common knowledge, recognizing frequently used synonyms, and limited natural language understanding that will expand over time.

These capabilities are available today in all Dataverse environments in the public cloud with English as their base language, and relevance search enabled. No additional configuration is required. We plan to support additional languages in the future.

You can help improve these relevance search capabilities by providing feedback through this option on the search results page:

And by enabling the optional “Improve Relevance Search results” setting located in the Power Platform admin center (Settings > Product > Features):

Review the documentation for more information about relevance search. If you are enabling relevance search for the first time, you can review how to configure searchable data here.