Result improvement rules
Define rules with configurable trigger conditions to boost, block, or promote search results for specific searches. You can also boost search results for documents matching elements of the user context, such as a user's country, OS, or device type.
Result improvement rules with boost actions override the default search result order produced by the AI Search machine learning relevancy feature. Rules with block or promote actions override the default result set for a search query.
Rule structure
- A set of conditions that determine when the rule activates.
- A set of boost, block, or promote actions that determine how the rule modifies the search query's results when activated. The rule can only include actions of a single type, but can include multiple actions of that type.
- A date range bounded by start and end dates. The rule can only activate during the specified date range.
- A language. The rule can only activate if this language matches the search query's language.
- Optionally, a set of one or more trigger conditions. Triggers can operate on individual search query keywords, the full search query, or elements of the search user's context such as their department, language, or browser version. The rule can only activate if all of its trigger conditions are met.
- Boost: Increase the relevancy scores for search results from an indexed source that satisfy your specified conditions. You can select any of the following options:
- Boost relevancy for all search results from the indexed source.
- Boost relevancy for search results from the indexed source that include field values matching a dynamically populated value from the user context.
- Boost relevancy for search results from the indexed source that include field values matching a static string value that you specify.
- Boost relevancy for search results from the indexed source that include field values matching synonyms from one of your synonym dictionaries.
- Block: Exclude specific results from the search query's results list.
- Promote: Display specific results at the start of the search query's results list.
Block actions take precedence over other actions. For example, if two result improvement rules block and promote the same search result, AI Search excludes the search result from the search query's results list.
Language dependence
Result improvement rule activation is language-dependent. You specify a language for each result improvement rule.
AI Search supports result improvement rules for all languages activated in your instance. For the list of languages you can activate, see Activate a language.
Interaction with other search features
The following table describes interactions between result improvement rules with Query triggers and other search features.
| Feature | Interaction with result improvement rules |
|---|---|
| Lemma and Unicode normalization | A search query term added by lemma or Unicode normalization can trigger a result improvement rule if it matches the rule's Query trigger. |
| Stop words | If a search query term that matches a result improvement rule's Query trigger is defined as a stop word, AI Search evaluates the rule trigger before removing the stop word from the search query. |
| Synonyms | By default, a synonym expansion term can activate a result improvement rule if it matches the rule's Query trigger. You can prevent synonym expansion terms from activating a rule with a Query trigger by clearing the Expand Synonyms For Query Conditions option that appears when you define the Query trigger. |
| Typo handling | An auto-corrected search query term can trigger a result improvement rule if it matches the rule's Query trigger. |