A guide to display ServiceNow Knowledge articles in Copilot for Microsoft 365

Introduction

This article will discuss the Microsoft Graph Connector for connecting to ServiceNow and indexing Knowledge articles in the Microsoft 365 Search Index and using Semantic Index these data are shown as prompt responses to user queries in Copilot for Microsoft 365.

As we are familiar, ServiceNow Knowledge Management is a module within the ServiceNow platform that serves as a library of known issues or potential issue resolutions. It allows users to create, edit, and view knowledge articles to share information across the organization.These articles contain descriptions of issues, their root causes, and resolutions. Knowledge Management in ServiceNow also enables the creation of multiple knowledge bases, each managed by one or more knowledge managers, and provides features for creating, editing, publishing, and retiring knowledge articles.

Microsoft Graph connectors are focused on indexing content from various data sources like ServiceNow to make it searchable and accessible within Microsoft 365 services. Graph connectors work by creating a connection to the data source, registering a schema that defines the structure of the data, and then ingesting content into the Microsoft Graph. This process enables Copilot for Microsoft 365 to surface and utilize the indexed data from ServiceNow and other sources in its responses, enhancing the discoverability and engagement of enterprise data within Microsoft 365.

Semantic Index for Copilot

The Semantic Index for Copilot is an improvement to Microsoft 365 search that helps users find information more quickly and easily within their organization's Microsoft 365 data. It does this by creating two indexes: a user-level index and a tenant-level index. Using text-based content that you create or interact with is included, such as emails, documents you're mentioned in, comment on, or share, User-level indexing creates a customized index for your eyes only on a working set of data, making it easier to use in everyday tasks. The tenant-level index adds text-based SharePoint Online files that are accessible by two or more people in your organization via site inheritance. The Semantic Index uses these indexes to understand the relationships between different words and concepts and to identify relevant information that may not be found using traditional keyword search methods. It allows organizations to search through billions of vectors (mathematical representations of features or attributes) and return related results. When a user issues a query either directly via search or through Microsoft Copilot, the semantic index is always in the security context of the user, and only content that the user is authorized to see is returned.

Workflow of Microsoft Copilot with Data Processing

Image Source: https://learn.microsoft.com/en-us/microsoftsearch/semantic-index-for-Copilot#data-flows

Step 1. User Prompt

• Initiated in Microsoft 365 apps.
• Sent to Copilot securely via HTTPS.

Step 2. Processing by Copilot

• Accesses Microsoft Graph: Retrieves user context and relevant information.
• Semantic Index Lookup: Analyzes personal and organizational data for deeper understanding.
• Modifies User Prompt: Refines the prompt based on gathered context.

Step 3. Response Generation

• Sends Modified Prompt: Forwards it to the Large Language Model (LLM).
• Receives LLM Response: Processes the text generated by the LLM.

Step 4. Post-Processing by Copilot

• Accesses Microsoft Graph: Utilizes context for further refinement.
• Semantic Index Lookup: Enhances response with relevant insights.

Step 5. Output

• Sends Final Response: Delivers the refined text back to the app.

Configure Microsoft Graph Connector

The following steps describe how to Index ServiceNow Knowledge articles in Copilot for Microsoft 365 using the Microsoft Graph connector:

Setting up a Microsoft Graph Connector is a three-step process. :

• Connect
• Configure
• Customize

To configure the Microsoft Graph connector, navigate to the Admin Center.

1. In the settings link in the left navigation, click the "Search and Intelligence" option. Once the details are visible, click on the "Data Sources" tab and then click on the "Add" option to add a new connector as shown in the below screenshot.

2. Connect: In the Other Data sources section, click on the ServiceNow Icon from the list of data sources and select "ServiceNow Knowledge". Click Next.

3. Create Connection: In order to create a connection, three fields need to be filled in:

• "Connection ID" - Connector ID should exclude the word "Microsoft". It must contain only alphanumeric characters and be a maximum of 32 characters. The connection ID creates implicit properties for your connector. This ID will be used when you make Microsoft Search API calls.
• Display Name - A user-friendly name makes it easier to identify the connector.
• Icon - It is not mandatory to use the default icon. You may substitute another image of your choice.

Select the mandatory notice Checkbox to create a connection.

4. Data Source Settings: Provide the ServiceNow instance URL. There are three options for authentication. Based on business needs and In case ServiceNow has the Microsoft Entra ID authentication you can directly map the Microsoft Entra ID objects (Individual users or O365 Groups) security needs and select the appropriate authentication type. Click on "Sign in" to connect to the ServiceNow data source.

5. Manage Search Permissions: The next step is to assign permission to each indexed content item. In case ServiceNow has the Microsoft Entra ID authentication you can directly map the Microsoft Entra ID objects (Individual users or O365 Groups). You also have the option to select "Everyone", selecting this option will provide access to all the users in the Organization.

Additionally, there is an option called "Quick Publish" that publishes connectors with default values. For the granular Advanced Settings click on "Save and Proceed".

6. Index Property Selection: In this section, Index properties are selected. By default, a few properties are pre-selected, based on the business needs either adding or removing the properties to the Index.

There is also an option to filter the data to be indexed. You can copy the filter from the ServiceNow View and paste it here to filter the relevant content.

7. Map Identities: Based on the settings in step 5, configuration options are available here. There are options to write a regular expression formula to map a non-MEID to a property in ME ID (Microsoft Entra ID) to apply the relevant permissions.

8. Assign Property Labels: To enable the search experience, assign the required source properties to each label. Some property labels are optional but can improve search relevance and ensure more accurate search results.

9. Manage Schema: A search schema provides options for Querying, Searching, Retrieve, and Refining. Defining a search schema allows search engines to determine what results appear on search results pages and what information end users can access. Users can perform search operations like refining, and querying using this configuration.

There are significant performance issues when search results are rendered with text content properties such as Page Content (Text). The configuration of these properties needs to be carefully considered.

10. Crawling the Content:  There are two options to select the crawling of the content. - Full crawl and Incremental crawl.

Important considerations

• Leaving start time empty for daily crawls: Connector picks a time based on your last crawl.
• No start time specified: Best practice to let the connector decide.
• The mentioned start time may be delayed by an hour due to network load.
• Overlapping crawls: Ongoing crawl continues, queued crawl starts only if it's a different type (e.g., full vs. incremental).

11. Once the connector is set up successfully, the crawling is triggered, and content is indexed and visible in the below Screen.

The "Item Indexed" column displays the indexed content.

12. In case any edits are required, you can navigate to the "Customizations" tab. On the left navigation click on the "Schema" option to 'Edit source properties, 'Add a new source property' and 'Update the semantic labels'.

13. Once the Search Content is indexed and the Semantic index is updated. Users can request information using Copilot and ServiceNow knowledge base articles will appear as shown in the below GIF (Network conditions may delay the rendering of the GIF image).

Image Source: https://www.youtube.com/watch?v=17rAOh9313g

Summary

The Microsoft Graph connector for ServiceNow makes ServiceNow Knowledgebase articles discoverable and accessible in Copilot for Microsoft 365 when you create a connection and configure it. If you wish to delve deeper into the topic, you can refer to the Microsoft documentation for a more detailed guide on building and configuring Microsoft Graph connectors for Copilot in Microsoft 365 in the references below.

References

• Overview - https://learn.microsoft.com/en-us/microsoftsearch/connectors-overview
• Semantic Index for Copilot: https://learn.microsoft.com/en-us/microsoftsearch/semantic-index-for-Copilot
• Licensing Requirements and pricing - https://learn.microsoft.com/en-us/microsoftsearch/licensing
• Can I use my Microsoft 365 Developer Program instant sandbox tenant to develop with Microsoft Copilot for Microsoft 365?
• https://learn.microsoft.com/en-us/microsoft-365-Copilot/extensibility/prerequisites#can-i-use-my-microsoft-365-developer-program-instant-sandbox-tenant-to-develop-with-microsoft-Copilot-for-microsoft-365