Once the Artifacts are added to Luma Knowledge, you can search and view the artifacts through any Luma-supported channels. Artifacts can be searched manually by the end-user using a simple phrase/question or via a request from a supported application, driven by the user action such as reporting an issue or submitting a service request.
Luma Knowledge uses metadata to classify and locate Knowledge Artifacts. Metadata identified during Artifact creation is matched against the metadata from the user query. This information is used to find relevant artifacts from the Knowledge Base.
How does Search Work?
There are various factors that impact a Knowledge Search result. While the factors such as the question, metadata impact the artifacts that are filtered as Best Response, tenant configurations determine the presentation of the result set.
Factors for Artifact Search
An Artifact appearing in a search result depends on the following:
Search Widget:
Metadata for User Query: Luma Knowledge generates metadata from the user’s query and uses it to find relevant artifacts from the Knowledge Base. The query’s metadata is used to find the Artifacts with the matching metadata. The search queries should be well-structured so that Luma Knowledge can extract appropriate metadata from the query and identify matching Artifacts.
Relevancy Score Range Percentage: The Relevancy score of an Artifact determines how relevant is the Knowledge available in the Artifact to the user’s question. This is a relative score and is calculated based on the best matching Artifact. All the Artifacts presented to the user in the result set, fall in a specific Relevancy score range that is derived from the Score Range percentage configured for your tenant.
For example, consider Score Range percentage for your tenant is set to 60%. If the highest Relevancy score of an Artifact for your search is 150, the system calculates the lowest permissible Relevancy score as 90 (that is 60% of 150). This means the artifacts that fall between the range of 90 and 150 are considered relevant and presented as search results.
Note: Relevancy score is used only in searches through Search Widget in Luma Knowledge or any other integrated systems such as ISM.
Luma Virtual Agent:
Confidence Score: Confidence Score is one of the factors that determine the result display in Luma-supported channels. The Confidence Score of an Artifact indicates how confident the system is that the information answers the user’s query. The score is calculated based on the metadata match.
Each Metadata type has a predefined confidence score assigned. When a user searches for Artifact through Luma VA, the system generates metadata from the user query. The query metadata is matched with the metadata of the Artifacts available in Knowledge Base to filter the relevant artifacts. If at least one phrase/value in a metadata type is matched, the respective score is awarded to the Artifact. More the matching metadata types, the better is the confidence score. For more details on calculating Confidence score, refer to /wiki/spaces/KNOWLEDGEMANAGEMENT/pages/26049413514.
Following are the scores assigned to each metadata type:Topic → 40
Path → 35
Subject → 30
Action → 25
Motivation → 5
For example, if the query metadata matches at least one phrase/value in each metadata type of the Artifact, it is considered a perfect match to the query with a confidence score of 100.
Confidence Score Threshold: Based on the Confidence Score and Confidence Score Thresholds configured for your Luma Knowledge tenant, the artifacts in the result set are classified in Confidence Levels. Luma VA uses the Confidence Level to decide if the user should be presented with Knowledge Articles, pre-defined service, or Both. For more information on Searching Knowledge through Luma VA, refer to Integration with Luma Knowledge.
You can customize the following Confidence Score Thresholds for your tenant to define Low and Medium Confidence Levels.Upper Confidence Score Threshold value for MODERATE Confidence level: This is the Confidence Score threshold for Confidence level MODERATE. All Artifacts with confidence score values between the defined value and Low Confidence Level are considered to have Confidence level -MODERATE.
Upper Confidence Score Threshold value for LOW Confidence level: This is the Confidence Score threshold for Confidence level LOW. All Artifacts at or under the defined confidence score value are considered to have a LOW Confidence level.
User Intent: User intent is derived from the search phrase. It indicates if the user wants to view Knowledge Artifact that answers the query or want to execute a predefined service to action the request.
For example:
“How to order a laptop” indicates that the user wants to view Knowledge Artifacts on how they can reset the password.
“Create a Service Request” indicates that the user wants help with a service that can reset the password.Matching Services on Luma VA: The user’s search request is also used to search predefined services available in Luma VA. Based on the Luma tenant configurations, the Virtual Agent identifies matching services and calculates the skill(s) confidence level. Services are presented to the users along with the Knowledge Artifacts if matching services with High or Moderate Confidence Levels are found. For more information on identifying Skills in Luma VA, refer to NLP Settings.
Result Presentation preference in Luma VA: In Luma VA, administrators can customize the way Knowledge and Services are delivered to the end-user. You can configure Luma VA to always deliver Knowledge followed by Skill, Knowledge & Service together or use the Decision Engine to derive a presentation based on the confidence scores and user’s tenant. For more information, Knowledge & Services Settings.
Tenant Configurations for Result Presentation
Following are the Tenant configurations in Luma Knowledge that determine the way identified Artifacts are presented to the end-user:
Maximum Artifacts per Query: This is the maximum number of artifacts that can be displayed for a Knowledge Search.
In case more artifacts are identified for the search, the system prompts that too many answers are found and displays the top relevant artifacts. The artifacts are sorted in descending order of their relevancy score.Maximum decision levels: The configuration determines the decision levels to be presented to the end-user before they can access the required Knowledge Artifact. When Luma Knowledge identifies multiple Artifacts related to different topics as a result set of a user request, a dynamic ‘Guided Conversation’ is generated to help the end-user find relevant information. You can customize the number of questions or topics that can be presented to the end-user, to gather additional information required to guide the end-user to the best matching Artifact.
Set the Maximum decision levels to 1 to always present a list of Artifacts.Maximum Topics per Decision Level: When presenting a Guided Conversations to the end-user, you can customize the number of options or Topics available at each decision level. Set the count high to create Guided Conversations.
Maximum Artifacts per Topic: The count represents the number of Artifacts that can be displayed under each Topic in a Guided search.