Published at: 2026-10-15

Import System Data

Import existing system data from the frontend and bring it into the knowledge base in a unified way.

Overview

Importing supports network disk, training assistant, sharing, knowledge article object, and CRM object data.

Before You Start

❗
The current account needs the frontend system data import permission and must be able to view the source data.
  • Confirm the target source directory, course category, sharing topic, or business object.
  • Confirm the field mapping for title, content, attachments, author, and publish status.
  • If AI enhancement is needed, confirm the target fields that are allowed to be generated or overwritten.

Accessing System Data Import

  1. Enter the Knowledge Base homepage.
  2. Click Import and select System Data Import.
  3. In the source list, choose Network Disk, Training Assistant, Sharing, Knowledge Article Object, or CRM Object.
Import panel with file, system data and third-party options
Except for CRM objects, each source page provides Ingestion Log and More Operations after entry. The configuration page usually contains a sync scope or data scope, field mapping, AI enhancement, Submit, Save and Re-ingest, and Cancel.
System data import panel overview

Source Permissions and the Default Space

  • System source data is carried by the source containers in the default knowledge space. The source folders actually shown depend on the current enterprise page.
  • Administrators need to be able to view the corresponding source data and have Edit or Manage permission on the target source folder.
  • Source system permissions and knowledge base resource permissions together determine the final visibility scope. Import rules cannot expand the authorization scope of the original source content.
  • You can also enter Ingestion Settings from the corresponding source folder in the default knowledge space. If the entry is not shown on the page, check the source folder permissions first.

Importing Network Disk Data

Converts files in a specified network disk directory into knowledge.

Setting Sync Directories and Ingestion Rules

  1. In the system data sources, select Network Disk.
Cloud Drive selected in the system data import list
  1. In Directories Available for Sync, find the directories to import and move them to Directories in Sync. Multiple specific directories can be selected.
Ingestion settings with sync directory lists and field mapping
  1. Check the directory hierarchy: after a parent directory is selected, its subdirectories are covered by the parent scope and become individually unselectable on the page.
  2. In the field mapping area, map network disk fields to knowledge fields. The mappings shown on the current page include:
Network disk field Knowledge field
File directory Knowledge category
File name Title
Empty source field Knowledge summary, can be AI-generated
Empty source field Knowledge outline, can be AI-generated
File Attachment content
Fixed value “Published” Publish status
Creator Author

Saving the Configuration

  1. To generate a knowledge summary or outline, enable AI Enhancement, and check the template and target fields.
  2. Click Submit to save the rule. Newly added data will follow this configuration.
  3. If you modify an existing rule and need to process historical data with the new rule, click Save and Re-ingest. This performs a full update of the network disk data in the sync scope.
  4. Check the processing results in the Ingestion Log.

Importing Training Assistant Data

Converts courses and their content under a specified course category into knowledge.

Choosing Course Categories and Configuring Fields

  1. In the system data sources, select Training Assistant.
Training Assistant selected in the system data import list
  1. From Course Categories Available for Sync, choose categories and move them to Course Categories in Sync. Multiple categories can be selected.
  2. When a parent course category is selected, all its subcategories are selected automatically. Before saving, check whether any subcategory you do not need is included.
  3. Configure the field mapping. The mappings shown on the current page include:
Training assistant field Knowledge field
Course category Knowledge category
Course name Title
Course summary Knowledge summary
Empty source field Knowledge outline, can be AI-generated
Course content fields such as audio/video transcripts Attachment content
Course instructor Author
Fixed value “Published” Publish status

Saving the Configuration

  1. To generate a knowledge outline, enable AI Enhancement and check the target fields.
  2. Click Submit to save; for existing rules that need to process historical courses, click Save and Re-ingest.
  3. Check the processing results in the Ingestion Log.
Course content may include instructor information, courseware, and audio/video transcripts. After the import completes, spot-check the accuracy of professional terms, numbers, and key steps.

Importing Sharing Data

Converts sharing content that contains a specified topic into knowledge.

Choosing Sharing Topics and Configuring Fields

  1. In the system data sources, select Sharing.
Post selected in the system data import list
  1. From Topics Available for Sync, choose topics and move them to Topics in Sync.
  2. Configure the field mapping. The mappings shown on the current page include:
Sharing field Knowledge field
Topic name Knowledge category
Content Title
Empty source field Knowledge summary, can be AI-generated
Content Rich content
Attachment Attachment content
Creator Author
Fixed value “Published” Publish status
Only sharing whose content contains a selected topic generates the corresponding knowledge entry.

Saving the Configuration

  1. Check that the synced topics are accurate.
  2. To generate a knowledge summary, enable AI Enhancement and check the target fields.
  3. Click Submit; for existing rules that need to process historical sharing, click Save and Re-ingest.
  4. Check the processing results in the Ingestion Log.

Importing Knowledge Article Object Data

Converts records in the existing knowledge article object into knowledge based on filter conditions and field mapping.

Setting the Data Scope

  1. In the system data sources, select Knowledge Article Object.
Knowledge Article Object selected in the system data import list
  1. The business object on the current page is fixed as KnowledgeArticleObj.
  2. In Data Scope, choose:
    • All: no additional filter conditions.
    • Filter by Conditions: only records meeting the conditions are processed.
  3. After selecting Filter by Conditions, set the field, comparison condition, and value. The page supports combining multiple conditions with AND or OR.
  4. Re-check the filter logic before saving. The example rule visible on the current page (“Publish status = Published”) is only a configuration example, not a mandatory condition for all enterprises.
  5. Configure the field mapping. The mappings shown on the current page include:
Knowledge article object field Knowledge field
Knowledge category Knowledge category
Title Title
Knowledge summary Knowledge summary
Rich content Rich content
Attachment content Attachment content
Publish status Publish status
Author Author
Editor Editor
Validity period setting Validity period setting
Validity start time Validity start time
Validity end time Validity end time
Valid or not Valid or not
Customer name Dedicated customer
The page also contains an enum-value mapping for “Internal dedicated knowledge or not”. Check each source value and target value one by one as shown on the current page.

Saving the Configuration

  1. Check the data scope and the field mappings for title, content, attachment, status, validity period, and dedicated customer.
  2. If AI enhancement is needed, check the enhanced fields and prompt.
  3. Click Submit; for existing rules that need to process historical records, click Save and Re-ingest.
  4. Check the processing results in the Ingestion Log.

Importing CRM Object Data

Converts records of any selected CRM business object into knowledge. Each object uses an independent ingestion rule.

Adding a CRM Object Rule

  1. In the system data sources, select CRM Object.
CRM Object selected in the system data import list
  1. Click Add Object and choose the CRM object to import in Business Object.
  2. In Data Scope, choose All or Filter by Conditions.
  3. If Filter by Conditions is selected, set the filter field, comparison condition, and value. The page supports combining multiple conditions with AND or OR.

Setting Field Mappings

The CRM object fields are determined by the selected object, so a fixed mapping set cannot be used. When configuring, map the object fields to the AI knowledge entry fields on the right.
CRM ingestion settings with business object dropdown
The page supports clicking Add Field to add more mappings, and also supports fixed-value mapping, for example mapping the fixed value “Published” to the publish status. In practice:
  • For the title, choose an object field that stably identifies the record content.
  • For the knowledge summary, rich content, and attachment content, choose source fields based on the business content.
  • Check the sources of the publish status and author; when a fixed value is needed, choose the fixed-value mapping on the page.
CRM ingestion settings with field mapping expanded
If content such as a summary needs to be generated or rewritten, enable AI Enhancement and check the enhancement target fields.
Click Submit to save the rule for this object.

Managing Existing CRM Object Rules

Return to the object rule list and operate according to the rule status:
  • Running: edit the sync rule, manually update, or deactivate it. The delete button is unavailable.
  • Not Running: edit the sync rule, enable it, or delete it. The manual-update button is unavailable.
CRM object rule list
When editing an existing rule, you can click Submit to save, or click Save and Re-ingest to reprocess historical data with the new rule.
The objects, fields, and filter conditions all depend on the enterprise’s CRM configuration. This article does not list fixed object names or fixed business fields, to avoid writing one tenant’s custom objects as universal capabilities.
CRM object ingestion converts records that meet the rule into knowledge. When the Agent queries CRM data in real time during a conversation, it still reads the current business data according to object permissions and record permissions. The usage and permission checks of the two capabilities are different.

Configuring AI Enhancement

Each source can enable AI Enhancement as needed. When configuring:
⚠️
AI-generated content directly overwrites the content synced from the original field mapping to the same target fields. Confirm the enhancement target fields before enabling.
  1. Turn on the AI Enhancement switch.
  2. Select or check the enhancement template for the current source.
  3. If a custom prompt is used, make sure the prompt contains:
    • the source field content to be learned, and
    • the target field to be enhanced.
  4. Confirm again that the target fields can be overwritten by AI-generated content.
  5. After saving the rule, spot-check the AI-generated content through the ingestion results.
AI enhancement enabled with prompt template selection

Difference Between Submit and Save and Re-ingest

Operation When to use Result
Submit Create a new rule, or only save the current configuration Saves the current rule; subsequent data follows it. After returning to the list, check the rule status
Save and Re-ingest The scope, field mapping, or AI enhancement of an existing rule changed, and historical data also needs to be reprocessed with the new rule Saves the rule and reprocesses historical data
Manual update A CRM object rule is running and an update needs to be actively initiated Initiates the update; check the processing result together with the rule status, and also check the log if the page provides a log entry
Re-ingesting may regenerate mapped fields or AI-enhanced fields. Before running it, confirm that the new rule will not overwrite knowledge content that must be kept.

Expected Results and Verification

After saving a rule, check at least the following:
  • Scope is correct: the directory, course category, sharing topic, or object filter conditions hit the expected data.
  • Title and content are correct: the title, summary, and rich content come from the expected fields.
  • Attachments are complete: files, courseware, audio/video content, or CRM attachments enter the attachment content.
  • Category and status are correct: knowledge category, publish status, validity period, and dedicated customer attributes match the rule.
  • AI content is usable: after AI enhancement is enabled, check the generated results in the overwritten fields.
  • Logs are normal: for sources that provide the Ingestion Log entry, view records that failed or were not processed as expected.

FAQ

Issue Resolution
Source data cannot be found Check the source system permissions and the target folder resource permissions.
Import scope is incorrect Re-check the sync directories, categories, topics, or object filter conditions.
AI content overwrites the original fields Disable the corresponding enhancement item, or re-select the target fields allowed to be overwritten.
Import task fails View the failure reason in the ingestion log or the Task Center.
Submit Feedback