You can use the "Query Builder" in the MaiAgent service to graphically create query metadata. Here's an example of creating contact query metadata:
Step 1: Enter the Contacts Area in the MaiAgent Backend Service
Contacts Area Entry
Step 2: Configure the Desired Contact
Find the contact that needs configuration and click Edit
Contact Edit Entry
After clicking, the following screen will appear
Query Builder Entry
Click the Query Builder directly to enter the graphical interface
Step 3: Select the Permissions You Want to Grant
Check the knowledge bases you want to grant access to, or optionally select specific documents within that knowledge base
Select Knowledge Bases and Documents
If you want to grant access to documents with specific tags, click "Show Tag Filter" to open the tag configuration space
Tag Filter Configuration
Here you can select the criteria you want to create, choosing "OR" (meets any condition) or "AND" (meets all conditions) as the document access standard within the granted knowledge bases
Message Configuration
You can also use the right-side menu shown below to select the knowledge permissions you want to grant:
Internal Conversation: Select Available Knowledge Permissions
You can also use "Show Tag Filter" to configure tag filtering conditions
Before sending each different message, you can still modify the right-side permissions for knowledge management
Available Documents List
Selected Documents Status
Currently able to reference three files: CV, SOP, and Survey
After unchecking CV in the next message and sending again:
Example of Unchecking Document
The Agent will no longer be able to use the CV file, and the citation library will not display CV citations:
Citation Library Change: After Removing CV
Permissions can be adjusted before sending each message, affecting the content that can be cited in that round's response.
Pre-configuration When Loading Web Chat
MaiAgent provides the ability to pre-configure query metadata before Web Chat formally initializes. If you want to restrict conversation content before contact (e.g., when providing Web Chat services to customers without registered contact accounts), you can provide query metadata during embedding, and the system will enable message-level knowledge base document filtering mechanism by default.