Boost Weights for AI Search Result Improvement Rules There are several actions in Result Improvement Rules. The two we will review here are Promote and Boost actions. Promote actions 'pin' one or more results to the top of the result set, based on the trigger condition. This means that regardless of what AI Search calculates is relevant, the promoted result(s) will appear ahead of all other results. For example, a common keyword may be 'laptop.' With a trigger condition of 'Query contains laptop,' an administrator can promote the organization's preferred and recommended laptop. Boost actions influence result order, either with a static field/value specification or based on user context. For example, instead of simply promoting a single result, you may want to nudge a group of results based on field values found on those results. In the same case, we have a set of Catalog Items that have the word 'Approved' in their name for the recommended items. This Boost action will increase the relevance of 'approved' items in your catalog, assuming your catalog item name contains the word 'Approved' and that the item would normally appear in your result set. Boost Weights add a scaling factor to the relevancy scores (how relevant the content is) of the results that would be returned for a search query. The following are example weights with the outcome: A Boost Weight of 1000 will increase the document's relevance by 100% A Boost Weight of 100 will increase the document's relevance by 10% A Boost Weight of –100 will decrease the document's relevance by 10%