Tenant Configurations

The article describes the various configurations available in Luma Knowledge that allow an administrator to customize how information is processed and presented to the end-user. These are tenant-level configurations that govern Curation, Search, Feedback, and other features in the system. These configurations should be set diligently as these configurations determine Dashboard Metrics calculations and values, Search Widget, Number of Artifacts to be displayed.

Navigate to Tenant → Configuration to view the following Configuration available for your tenant:

Feedback

Luma Knowledge captures Feedback from end-users is captured to improve Knowledge from time to time. Feedback configuration parameters facilitate setting up a conversation between the Luma Knowledge system and the end-user during the feedback process, time intervals to check up ineffective or obsolete artifacts to improve artifacts' efficacy. These parameters will also determine Dashboard metrics such as Feedback Response Rate, Volume, and Accuracy, User request insights. The following parameters are set up in Luma Knowledge to obtain feedback. Default values are given below, and you can set up values based on your needs. For example, Time Interval to check Negative Feedback for Artifact has been set to 360 days, and in case you want to monitor Artifact feedback more closely, the desired value can be set.

Item

Description

Default value

Item

Description

Default value

Artifact Feedback Question (Main)

The feedback question to be presented to the end-user to check if the artifact was useful.

Did the artifact answer the user’s question/problem (Yes or No)?

Artifact feedback first follow-up question

The first follow-up feedback question for the end-user after the user selects the feedback as not useful.

Did the artifact’s topic match the original knowledge request (Yes or No)?

Artifact feedback second followup question

The second follow-up feedback question for the end-user after the user selects the feedback is not useful.

Was the artifact understandable (Yes or No)?

Time Interval to check Obsolete Artifacts

Time interval (in days) to check for obsolete artifacts.

180

Time Interval to check Negative Feedback for Artifact

Time interval (in days) to check for ineffective artifacts.

180

Negative Feedback Count Threshold for Ineffective Artifacts

The minimum number of negative feedbacks to consider an artifact as ineffective.

5

Time Interval to check for Ineffective Best Responses

Time interval (in days) to check for ineffective best responses.

180

Feedback Count Threshold for Similar Ineffective Best Responses

The minimum number of Best Responses with negative feedback required for ineffective Best Response analysis.

5

Feedback response rate

Feedback response rate target for a channel

75

Search

Search configuration parameters determine the Search widget structure, the way Artifacts are to be displayed for a search query. Here you can set up the number of decision levels and topics per decision level to be displayed for search results.

Item

Description

Default value

Item

Description

Default value

Maximum Topics per Decision Level

The maximum number of topics allowed per decision level.

12

Score Range Percentage

The percentage of the Highest Score to determine the lowest eligible score for search results. Any document with a score lesser than the lowest eligible score will be discarded from Best Response.

60

Maximum Artifacts/Documents per Topic

A maximum number of artifacts are allowed per topic.

10

Maximum Artifacts per Query

 

Maximum number of catalogs to be returned for a search query.

30

Maximum Catalogs per Query

A maximum number of artifacts are allowed per topic.

30

Minimum Catalogs for guided result

 

Minimum number of catalogs required to show guided navigation

10

Upper Confidence Score Threshold value for LOW Confidence level

Upper threshold value (inclusive) of confidence score for Confidence level LOW. Any search result having confidence score value between 0 and best.response.confidence.band.low.upper.threshold (inclusive), will be considered having Confidence level as LOW

0.5

Upper Confidence Score Threshold value for MODERATE Confidence level

Upper threshold value (inclusive) of confidence score for Confidence level MODERATE. Any search result having confidence score value between best.response.confidence.band.low.upper.threshold (exclusive) and best.response.confidence.band.moderate.upper.threshold (inclusive), will be considered having Confidence level as MODERATE

0.75

Enable curators to search in latest artifacts

The flag enables the curators to view and search the latest artifacts (in the Draft state) from integrated applications like Luma Virtual Agent, ISM.

No

Exclude Articles/Catalogs with Low confidence from search results

The configuration enables the curator to exclude Artifacts with Low Confidence level from the search results. Only the Artifacts with High and Medium confidence levels are presented to the user.
The configuration removes the Artifacts with Low Confidence level from the result set displayed in Luma Virtual Agent as well as the Luma Knowledge Search Widget.

No

Exclude Articles/Catalogs with Moderate confidence from search results

The configuration enables the curator to exclude Artifacts with Low and Medium Confidence levels from the search results. Only the Artifacts with High confidence level are presented to the user.
The configuration removes the Artifacts with Medium and Low Confidence levels from the result set displayed in Luma Virtual Agent as well as the Luma Knowledge Search Widget.

No

Confidence score weightage calculation

The configuration is used to determine the confidence score weightage calculation. With 'Static', the confidence score weightage assigned to each Metadata Type is fixed and the Artifacts Confidence score is evaluated accordingly. 'Dynamic' means the weightage is assigned based on Ontology generated from the User phrase.

Static

Artifacts

Here you can define the maximum allowed size for files, and multimedia attachments. Average Cost per Ticket is an important configuration parameter that determines dashboard metrics such as return on Investment (RoI) for a Channel, Results Summary.

Item

Description

Default value

Item

Description

Default value

Average Cost per Ticket (in USD)

The average cost of a ticket for the organization which is either deflected by LUMA Knowledge or created due to the unavailability of knowledge.

1

Multimedia Size Max Limit

Maximum allowed size for a multimedia attachment.

5

Create artifact minor version

Flag to create minor versions of an artifact upon each update. Set to True or False.

False

Attachment Size Max Limit for docx format

Maximum upload size (in MB) for a word document (.docx format).

3

Attachment Size Max Limit for pdf format

Maximum upload size (in MB) for PDF file (.pdf format).

3

Attachment Size Max Limit for txt format

Maximum upload size (in MB) for text file (.txt format).

3

Attachment Size Max Limit for xlsx format

Maximum upload size (in MB) for the excel spreadsheet (.xlsx format).

3

Default Artifact state on curation

Indicates the Artifact state when it created in Luma Knowledge. Based on the selected configuration and identified metadata, a new artifact can either remain as draft or bypass the manual review process to get directly Published.

Following are available options:

  • Publish if basic metadata available: Artifact is automatically Published if the basic metadata, i.e. Topic, Subjects and Actions,

  • Publish Draft: Artifact is set to Draft state.

  • Publish Anyway: Artifact is automatically Published, with or without mandatory Artifacts.

Publish if basic metadata available

Enable Artifact QnA Generation

The configuration is used to enable or disable QnA Generation Service for the tenant. The QnA Service generates QnA pairs from the document.

Note: If disabled (set to No), the Create artifact → Documents option (to create artifact from a document) is not available to the curator.

Yes

General

Here you can set up the general configurations for your tenant.

Item

Description

Default value

Item

Description

Default value

Time zone

Time zone for the tenant.

GMT

Download Format

By selecting option, files will be downloaded in appropriate format

Excel

Auto Generate Synonyms (Vocabulary)

Enable the configuration to generate synonyms (vocabulary) from metadata automatically.

No

Scheduler

This section contains the scheduler configurations for your tenant.

Item

Description

Default value

Item

Description

Default value

Posterior Probability Feedback Threshold

Minimum number of feedbacks since the last calculation required to re-calculate the posterior probability.

Note: Posterior Probability is used to arrange and align Topics in the result set so that the most relevant artifact is quickly accessible to the end-user.

500

Bulk Import

This section contains the configurations required for the Artifact Bulk Import.

Item

Description

Default value

Item

Description

Default value

Bulk Import allowed file size

The maximum upload size (in MB) for zip file for bulk import

30

Maximum No.of files allowed

Maximum number of documents that can be imported in a single bulk import

50

Allowed file types

These are the allowed file types allowed in bulk import

.docx, .pdf, .xlsx

Attachment Size Max Limit for docx format

Maximum upload size (in MB) for a word document (.docx format).

2

Attachment Size Max Limit for pdf format

Maximum upload size (in MB) for PDF file (.pdf format).

2

Attachment Size Max Limit for xlsx format

Maximum upload size (in MB) for the excel spreadsheet (.xlsx format).

2

Use Mini Artifact for failed artifacts during import

This indicates that Mini Artifact template should be used as fallback option to reprocess any document that fails during bulk import because of template validation.

False

Skip topic generation

This indicates that the automatic topic generation from the document name, Summary and other Ontology fields during Artifact creation should be skipped.
Only the topic hierarchy provided in the document should be used to create topic and artifact association in Knowledge graph.

False

Language Model

This section contains the Language Model configurations for your tenant. Here you enable or disable Open AI features for your tenant.

Item

Description

Default value

Item

Description

Default value

Enable Open AI

Activate Open AI for your tenant to leverage advanced language models to effectively manage your knowledge base and conduct more efficient searches.
With Open AI, Luma Knowledge can summarize content, generate Key phrases,generate precise answers and resolve Knowledge Gaps

No

Open AI features

Enable the required Open AI functionalities for your tenant. Ensure that Open AI is already enabled for your tenant.

 

To update your tenant configurations:

  1. Locate the tenant configuration.

  2. Update the value.

  3. Click Save.