--- schema_version: "1" id: e5ec109a9c337de7b62ef13a206ff5e70472dde4bcf976637a3e1883bf2b57b6 title: Einstein Search Terms collection: admin version: current locale: en-us portal_url: https://docs.salesforce.com/admin/current/en-us/ai/264-0-0/search_glossary.md canonical_url: https://help.salesforce.com/s/articleView?id=ai.search_glossary.htm&release=264.0.0&type=5 source_site: help.salesforce.com content_type: documentation content_hash: 202ba79c9184de7b9446048f8b8dcacd23dc92f38d63f7aff58f652c7b7995c6 description: Learn the definitions of terms associated with Einstein Search and how they apply to the search experience. products: - Cross Cloud release: "264" --- # Einstein Search Terms Learn the definitions of terms associated with Einstein Search and how they apply to the search experience.

Term	Description
Lemmatization	Resolves a word to its root while retaining the meaning of the word. The root form helps to identify variations of a noun or verb. A search for the word `feature` returns results for features. A search that includes the word `adds` matches items that contain the words adds, adding, and added. A search for the word `name` can return the words name, names, and named. Matches on different forms of a word rank lower than matches on the exact term in search results. You can search on one version of a word and get matches for other forms of that word.
Synonym Groups	Words or phrases that the search engine treats equally. When users search, they find matches for all terms in a synonym group. For example, if a synonym group contains the terms thumb drive, flash drive, and memory stick, a search for `thumb drive` returns all three terms. The two types of synonym groups are: Standard synonym groups: Common English language groups provided by Salesforce and available to all orgs for Salesforce Knowledge entities. An admin must enable standard synonym groups. Custom synonym groups: An admin creates a custom group specifically for an org.
Spelling Correction	Search returns results that match corrected spellings of your search terms. A message at the top of the search results page tells you that results include spelling corrections. If results include matches for your original search term, search again to show results for only that term. Spelling correction is also available for instant results, although no message is shown. For instant results, spelling correction works only for one term, even if you enter multiple terms. If multiple terms are misspelled, correction is only applied to the first term. Instant results spell correction isn’t available for knowledge articles. Spelling correction is applied in these ways. Search uses the name fields of existing records to identify potential spell corrections. Sometimes search results from Global Search differ from searches on an object. This difference can be due to the size of the data set for the name fields for an object. With Global Search, you get more results, as there are more objects searched on for a term. For all objects except Chatter, searches for spell-corrected terms apply to all searchable fields of supported objects. For Chatter, we search the post and comment text. For each object, there can be a different spelling correction for the same term. Spelling correction isn’t performed on search queries that include a wildcard (jo*), a number (PD200), exact matches, or terms in quotation marks. For spell correction to work on a supported object with multiple name fields, all name fields must be accessible for spell correction to apply. If the object has only one name field, that field is of AutoNumber data type, search does not apply spell-correction to the object.
Nicknames	Can’t remember if someone goes by Mike or Michael? You can look for a name, and the returned results include matches for related nicknames. Nicknames are ranked lower than exact matches in search results, but not always. Many factors determine how results are prioritized. Examples of related nicknames that search returns: `Milk` returns as Mike, Mikhail, or Mikah. `Mike Smith` returns as Michael Smith. `Chelsey` returns as Chelsea. `Yung` returns as Young.
Expanded Searches	Salesforce expands searches by including records whose name contains part of the term that you entered. For example, when you search for `lap`, the term laptop is also returned. When your search contains punctuation or special characters, Salesforce expands the query to look for other possible results. For example, searching for `1234.Silver` returns a record for XYZ-1234-Silver.
Knowledge Article Query Fallback	If your knowledge article search finds fewer than 20 results that contain all your search terms, search returns results that match any of your search terms. This feature, called query fallback, is only for the Knowledge Article object. For example, when you enter the term `beach house`, you can receive one of these responses. Results without Query Fallback: Returns records with beach and house in the searchable fields. Results with Query Fallback: Returns records with beach or house in the searchable fields.

- **[Objects and Languages Affected](https://help.salesforce.com/s/articleView?id=ai.search_glossary_objects.htm&release=264.0.0&type=5)** Lemmatization, synonym groups, spelling correction, and nicknames all affect different objects. Review each term, and see which objects and languages it searches on.