Advanced Usage
You have completed the basics and now want your Digital Worker to understand your business. Build a semantic model so it can reason accurately over real business data, beyond answering questions from a Knowledge Base.
Advanced usage turns your Digital Worker from a knowledge assistant into a business expert for a specific position. Set it up in five steps: assign accounts → build semantics → create Skills → create a position → use it in the workspace.
Assign accounts and configure data sources
Before you begin, ensure accounts and permissions have been assigned by position under "Admin Console > Platform Management > User Management" and your enterprise data sources are connected. Then start building the position's semantic model.
Assign accounts (prerequisite)
Under "Admin Console > Platform Management > User Management", organize users and groups by position: platform administrators, frontline business users, and managers of specific positions. A system administrator can complete this in advance.
Enterprise use requires planned user permissions and connected enterprise data sources.
Assign accounts
Create accounts and configure access by position so each role has the appropriate permissions.
Admin Console > Platform Management > User Management
Configure users and groups by position in one place: platform administrators, frontline business users, and managers of specific positions.
Configure data sources
Connecting data sources is a prerequisite for enterprise semantic modeling. Object types, such as customers, and their properties, such as customer tiers, must map to physical database tables to support queries and calculations.
Admin Console > Platform Management > Data Source Configuration
Build an ontology (semantic modeling)
Ontology building is central to enterprise use. Use "AI Generation" to create a semantic model and significantly reduce modeling time.
Admin Console > Semantic Center > AI Generation
In the Semantic Center, click "New Graph" to create a semantic graph space, select relevant data sources, and choose "AI Generation" as the initialization method.
Describe the business scenario and operational goals in detail, select the relevant physical tables, and click Generate. AI extracts core business entities, creates object types and Business Logic, identifies relationships, and configures physical table mappings.
Review the added, modified, and deleted definitions on the current page. Check that objects are complete, properties are accurate, and relationships are appropriate.
Correct any unsuitable definitions manually. For bulk changes, use YAML editing mode to edit the semantic model source directly.
Create Skills
Equip each digital position with dedicated Skills for generating documents, sending messages, and calling APIs, so your Digital Worker can take action as well as answer questions.
Admin Console > Skill Center
Official built-in Skills
The platform includes general-purpose and sales Skills that are ready to use:
| Category | Built-in Skills |
|---|---|
| General office work | Word documents, PPT presentations, Excel spreadsheets, PDFs, illustrated briefs, and more |
| Collaboration | Follow-up emails, meeting summaries, and more |
| Sales | Sales collateral generation, customer outreach strategies, lead discovery, and more |
Enterprise custom Skills
Create Skills for specific positions in either of two ways:
Skill Center > New Skill in the upper right
Describe the Skill in natural language, such as a Skill for improving sales scripts. The model generates Skill documentation from your requirements. Customize and debug it, then click "Save" → "Publish" in the upper right.
Upload a local .zip or .md file. The .zip root must contain SKILL.md, with a clear name and description defined in YAML. After importing, you can debug and publish the Skill.
Create a position
Configure semantic information for the position, including Object Graphs, Business Logic, Analysis Views, and Knowledge Bases, to give it business understanding. Add Skills to give it the ability to act. Together, these define a position for a specific business function.
Admin Console > Positions > Create Position
Configure basic information
| Setting | Description | Example |
|---|---|---|
| Basic information | Name, avatar, and description | Name: Product Selection Expert |
| Associated object types | Select the Object Graph scope the position can access | Customers, projects, charging stations, devices |
| Associated Business Logic | Select the business rules the position can use | Churn alerts, equipment inspection rules |
| Associated Analysis Views | Select the analytical perspectives the position can use | Customer analysis, operational efficiency |
| Skill scope | Specify which Skills the position can invoke | All general-purpose Skills, product selection, and competitive analysis |
| Knowledge Base associations | Link Knowledge Bases the position can reference | Product manual and industry report folders |
Manage permissions
Assign users and administrators to the position.
Use the workspace
After configuration, business users can sign in to the workspace and use Digital Workers configured for their positions. Switch Digital Workers in the chat area to suit each scenario.
Before rolling out to users, administrators should perform an acceptance check:
- Ask business questions the Digital Worker should answer, and confirm it references the Object Graph and Knowledge Base correctly.
- Ask questions outside its permissions that should be rejected, and confirm the boundaries in the position description take effect.
- Run 1–2 custom Skills and confirm the Output structure and content quality meet expectations.
What to read next
Try it: Follow these five steps to build a Digital Worker for a real position in your organization, and put it to work with your team over the next week.