1. Introduction
A mature ServiceNow Configuration Management Database (CMDB) would contain vast amounts of data related to hardware, software, services, etc. and their relationships. It is often challenging for users to find the information required from their organization’s CMDB. Querying this complex data model typically requires knowledge of the database schema and structure. This is where the CMDB intelligent search feature based on natural language queries (NLQ) can be a game changer. It enables users to find what they need from the CMDB using plain English questions and statements. So, there is no need for them to understand the underlying database semantics or write any complex queries.
2. Accessing and using CMDB Intelligent Search
Step 1: Navigate to CMDB intelligent search
From the ServiceNow banner, click on the ‘Workspaces’ menu, and then click on ‘CMDB Workspace’ to open it.
At the top portion of the ‘CMDB Workspace’, you can see ‘Intelligent search’ with a search box below that.
Step 2: View search tips (Optional)
Before entering the search keywords, users also have the option to see tips about how to perform the search more effectively. Click on the ‘Search tips’ URL near the search button to view some tips about how to improve the search results.
A search tips pop-up view appears, as indicated in the following screenshot.
The following 4 tabs are available on the search tips pop-up:
- Single-table search: Tips for searching a single CI class and its associated attributes. This will be displayed initially when the search tips pop-up opens.
- Multi-table search: Tips for multiple classes and related tables.
- Advanced filtering: Information on keywords that can be used as filters to narrow down the results.
- Relationships: Tips about how to query relationship data.
Step 3: Use natural language to search for the information which you need
The NLQ interface provides a simple input field (i.e., the search box). Users can simply type in what they're looking for using complete sentences in natural language in this search box. E.g., Show me all servers located in San Diego.
We can see sample searches and past searches displayed under the search box. ServiceNow also suggests potentially relevant table names and field names while typing.
Step 4: Refine the query if needed
After entering the initial search, ServiceNow will attempt to map the natural language to a formal database query. If any ambiguity remains, it will prompt you to further clarify aspects of the query through a dialog. For example, in the following screenshot, there is a pop-up that gives us the option to specify the table name for the CI class which we intend to query in our search.
Click the ‘View search results’ button after reviewing the information asked in the pop-up window and changing it if required.
Step 5: View and interact with the results
The search results will be displayed after clicking the ‘View search results’ button in the previous step. Here you will find options to export this data, provide feedback on the relevance of the results, view and modify the filter conditions, etc.
Note: If the constructed CMDB query contains more than one table, then the ‘View in Query Builder’ button is displayed with the search results. Click the button to open the CMDB Query Builder with your query fully constructed on the Query Builder canvas. You can use the Query Builder to continue editing the query.
3. Enhancing the accuracy of CMDB Intelligent search results
The NLQ capability comes pre-trained with a certain amount of common language understanding. However, its accuracy can be further enhanced by configuring the following:
Synonyms:
Synonyms in ServiceNow CMDB enhance the search functionality by allowing users to find records using different terms that have the same or similar meanings. We can use this to define pairs of equivalent terms. This feature ensures that searches are more flexible and inclusive, accommodating variations in terminology. For instance, if "laptop" is set as a synonym for "notebook," searching for either term will yield the same results, helping users find the information they need without worrying about using the exact terminology. This can be configured by navigating to NLQ>Synonyms and creating a new record with the required information.
The highlighted record in the screenshot below shows an example of the synonym ‘Owner’ being mapped to ‘Owned by’ field on cmdb_ci. This means that a search using the keyword owner will consider the ‘Owned by’ field while retrieving the search results.
The following screenshot shows the form view of a record available OOB on the Synonyms table. Here it indicates that a search involving any of the keywords mentioned in the ‘Synonyms’ field will consider the ‘Name’ of the ‘Model ID’ of the server while retrieving the search results.
Semantic Shortcuts:
This can be accessed by navigating to NLQ>Semantic Shortcuts. Semantic shortcuts help Natural Language Query (NLQ) systems in understanding various types of user requests. They map common words to database columns. But they only apply when certain conditions are met. Some semantic shortcuts are available OOB. Organizations should create additional custom semantic shortcuts by considering their requirements, use cases and the keywords used in past searches (this can be checked by navigating to NLQ>Logs)
As per the following example, a search mentioning ‘Servers which I own’ will consider records where the ‘Owned by’ field in the cmdb_ci_server table matches the current user who is performing the search.
CMDB Implicit Relationships:
You can enhance the Intelligent Search feature's ability to find relevant results by defining implicit relationships between CI classes. Implicit relationships can be particularly useful for queries involving service offerings and application services. This can be accessed by going to NLQ>Cmdb implicit relationships from the application navigator. The relationship tab of the ‘Search tips’ window also contains a link to access this. An implicit relationship defines the association between two tables and includes any filters to be applied.
4. Conclusion
The CMDB intelligent search feature helps ServiceNow users to get comprehensive information from the CMDB without needing to understand database complexities. As a result of enhancing the user experience while querying the CMDB, it can also indirectly enable to improve CMDB effectiveness as part of various use cases like service operations, compliance, planning, etc. Intelligent search empowers more stakeholders to capitalize on authoritative CMDB intelligence using natural conversation rather than rigid, structured queries.
