Current Problem:
In Confluence Automation, the search logic used for rule configuration is separate from the global search and does not respect the global "Indexing Language" setting. The current behavior is as follows:

When a user enters a single character, the system suggests results using prefix or partial matching.

When a user enters two characters, the system prioritizes exact or near-exact matches.

This leads to a situation where, if a two-character word (e.g., "電話") is not indexed as a standalone token, it may not appear in the search results. As a result, users may find that entering the first character returns results, but entering the full word does not, which is confusing and degrades the user experience.

Suggested Solution:

Review and enhance the search logic in Confluence Automation so that multi-character queries (especially two-character words) are matched more intuitively, even if the word is not indexed as a standalone token.

Consider aligning the search behavior with the global search or providing an option to respect the "Indexing Language" setting.

Ensure that users receive consistent and expected search results regardless of the number of characters entered.

Workaround:
Currently, users can sometimes find results by entering only the first character of a word, but this is not intuitive and does not solve the underlying issue for all cases.