Define Searchable Content for AI Search
Content such as records in tables needs to be defined and indexed to be surfaced to users. This step explains what indexing is and reviews the default configuration for Knowledge articles and Catalog items. We go through how to index additional tables, indexing settings and security.
We'll cover the steps for Indexing
The first step is to make data available to the search engine. This is done via a process named indexing. Indexing allows AI Search to parse records from the different sources and organize information before a search to enable super-fast responses to queries.
Each table and its child tables need one record in the Indexed Sources table to be indexed by AI Search. Navigate to AI Search > AI Search Index > Indexed Sources to see your Indexed Sources. By default, you should be able to see the Catalog Item Table, Knowledge Help Content and User Table Indexed Sources.
Anatomy of an Indexed Source
The condition prevents feeding outdated or irrelevant data to the search engine indexer. It should be a high-level filter, for example, based on active records. More granular filters are defined in the next step when configuring the search sources to create different experiences.
You can configure attributes and field settings to control the indexing behavior for source records. Attributes (Advanced Configuration related list) control the indexed source's behavior at the record level, while field settings (Field Settings & Mapping related list) define the behavior for individual fields. Examples of behaviors controlled by the attributes are the default indexing behavior for attachments or tags, or specifying whether to index translated fields. Learn more here. Examples of behavior controlled by field settings are indexing additional fields from the reference field by dot-walking or disabling text indexing or searchability for a field. Learn more here.
AI Search uses settings to automatically purge stale records from the index and optimize search performance. To limit the set of records indexed from source tables, retention policies are required for Task [task] table or tables that extend it. They are optional otherwise. Learn more here.
When a new indexed source is defined, AI Search automatically begins indexing records when they are changed (either create, update, or delete).
To index existing records from the source table, perform a full table index by following the procedure here. AI Search indexes about 10k events every 30 secs.
An external content schema table is used instead of a Now Platform table for external content. For more details on configuring indexed sources for external content, see Indexing and searching external content in AI Search.
Content Security in AI Search
AI Search filters search query results and displays only records that the currently logged-in user can access. Content security is automatically enabled and isn't configurable. When indexing records from Now Platform tables, AI Search preserves the following access control settings:
- Role-based field-level access control list rules (ACLs)
- Non-scripted conditional ACLs
- User criteria
- Before Query business rules
- Domain separation
AI Search doesn't support scripted field-level ACLs.
What's next?
After the content is indexed, you can make the search available for different applications.
Go back to the Quick Start Guide.
References
 |
Documentation |
 |
FAQ |
Don't forget to subscribe to our AI & Intelligence forum to get the latest updates!
