---
sourceDocument: Australia ServiceNow AI Platform Administration
sourceDocumentLink: https://servicenow-prod.fluidtopics.net/r/platform-administration
Release :
- australia
ft:locale :
- en-US
ft:publication_title :
- Australia ServiceNow AI Platform Administration
ft:clusterId :
- platadm
bundleId :
- platadm
workflow :
- Platform
---
# Machine learning relevancy
# Machine learning relevancy in AI Search {#ariaid-title1}
Release version: Australia
Updated March 12, 2026
 3 minutes to read
Summarize
 Summarized using AI
This content was generated using new OpenAI-powered functionality. Results are provided on an as is basis and are not guaranteed to be accurate or complete.
## Summary of Machine Learning Relevancy in AI Search
AI Search leverages machine learning to prioritize the most relevant search results based on user interactions.
This functionality is automatically enabled and not configurable, ensuring that search experiences continuously improve without manual adjustments.
Show full answer Show less
## Key Features
* **Relevancy Scoring:** Each search result is assigned a relevancy score specific to the document, search terms, and user, with higher scores appearing first.
* **Search Signals:** User interactions with search components generate signals that inform how relevancy models are tuned over time.
* **Tuning Process:** Every 30 days, AI Search evaluates and modifies relevancy models based on aggregated search signal data, employing A/B testing to ensure improvements.
* **Profile-Specific Models:** Each search profile has its own relevancy model, and changes in one profile do not impact others.
* **Auto-Complete Suggestions:** A separate relevancy model ranks suggestions based on freshness and title match, distinct from general search results.
* **Score Viewing:** Administrators can access relevancy scores through the Search Preview UI in the Advanced AI Search Management Tools.
## Key Outcomes
With effective machine learning relevancy tuning, ServiceNow customers can expect enhanced search results tailored to user behavior, leading to improved search efficiency. However, successful tuning requires a minimum of 10,000 stored search signals; otherwise, the original relevancy model remains in use. Additionally, updates to relevancy scoring may occur during version upgrades, with previous models potentially reverting to defaults if trained over two releases ago.
AI Search displays the most relevant search results for a query first. Machine learning automatically tunes search result relevancy scoring for search experiences based on aggregated user interactions.
Machine learning relevancy is automatically enabled and isn't configurable.
## Relevancy models and scoring {#machine-learning-relevancy-ais__section_avh_5y2_1nb}
AI Search uses a relevancy model to compute a relevancy score for each result returned by a search. Documents with higher relevancy scores appear first in the result set. A result's relevancy score is specific
to the particular document, search terms, and user associated with the query.
Each search profile includes its own relevancy model. You can't view, modify, or delete this relevancy model.
Note:
AI Search doesn't apply relevancy ranking to *** universal wildcard queries. Results from *** queries appear in an unspecified order.
## Search signals and machine learning relevancy tuning {#machine-learning-relevancy-ais__section_obm_cbf_1nb}
AI Search UX components record signals associated with user searches. These search signals include data on how search users interact with the search input field, auto-complete suggestions, facet
and source facet bucket filters, Genius Result answer cards, and search results. To learn more about how search signals are recorded and stored, see [Search signals](https://servicenow-prod.fluidtopics.net/alvwyMBkcEw_TcmcTjEvbw "Search stores signal data from user searches. The aggregated signal data helps improve search result relevancy in your system.").
Machine learning relevancy uses data from these search signals to intelligently tune relevancy models on a continual basis. Every 30 days, AI Search computes a new version of each relevancy model, iteratively modifying its parameters and regression testing it against aggregated search signal data for the search profile. When this tuning process is
complete, AI Search compares the existing and new relevancy models to see which one produces better matches for user search behavior as recorded in the historical signal data.
If the new relevancy model produces better results with the signal data, AI Search uses its modified parameter values to perform A/B testing evaluations of live search traffic for the search profile. These evaluations test individual parameter changes to verify that they produce better search relevancy.
Note:
For details on the search query parameter evaluation framework used to perform A/B testing evaluations, see [Search query parameter evaluation framework](https://servicenow-prod.fluidtopics.net/IepXMRoEzEqaTf_QG8tsNA#srch-qry-parm-eval-fw-ais "AI Search performs A/B test evaluations that compare result relevancy for alternate sets of search query parameter values. These evaluations determine the models that AI Search uses for machine learning relevancy and for Q&A Genius Results.").
If the new model outperforms the original model in both the historical search-match comparison and the A/B testing, AI Search sets it as the active relevancy model for the search profile, overwriting the existing relevancy model. The updated relevancy model remains in use until the next tuning cycle begins.
These relevancy model tuning processes occur separately for each search profile. Changes made to the relevancy model in one search profile don't affect relevancy models in other search profiles.
Note:
When you upgrade to Australia from a previous release, the default relevancy scores for your search results may change. Relevancy models trained in the previous release should continue to produce the same result ordering. Models trained more than one release ago may revert to the default relevancy model.
Successful relevancy tuning requires at least 10,000 stored search event signals for a search profile. If the search profile has fewer than 10,000 signals stored, relevancy tuning fails with an error message and AI Search continues to use the original relevancy model for that search profile.
## Relevancy model for auto-complete suggestions {#machine-learning-relevancy-ais__section_fpl_ljy_gtb}
AI Search uses a dedicated relevancy model to rank records for display as auto-complete suggestions in the search field. This relevancy model scores records based on their freshness and on search query term matches in their title fields. The system doesn't train this auto-complete suggestion relevancy model. For details on configuring auto-complete suggestions, see [Auto-complete suggestions in AI Search applications](https://servicenow-prod.fluidtopics.net/3webB35t96NRwaF09nqFeA#auto-complete-ais "Automatically display intelligent suggestions in your search application's input field, helping users formulate their searches or jump directly to relevant results. Suggestions can include relevant search results, popular queries, recent queries from the user's personal history, and popular search results.").
## Viewing relevancy scores for search results {#machine-learning-relevancy-ais__section_h4b_kb5_ffc}
Search administrators can view the scores for search results in the Search Preview UI from the Advanced AI Search Management Tools ServiceNow® Store application. For details on using this feature to investigate search behavior, see [Search Preview UI for AI Search](https://servicenow-prod.fluidtopics.net/BlC25Y7yfL4X1GpWLeM1uw#ai-search-preview "The Search Preview UI enables you to test AI Search queries using settings from a selected search profile. Administrator tools help you review search query performance data and feedback, debug queries, and override default query settings for testing purposes.").