What the data for this category looks like
Urban commercial bank marketing-related data comes primarily from internal core business systems, customer relationship management systems, and potential customer intent data collected at offline branches. Product data updates synchronously when products go live or parameters are adjusted. Customer intent data is generated in real time when forms are submitted.
Data is organized in two structural categories. The first is structured product metadata, including fields such as product identifier, investment threshold, and duration. Units are code, yuan, and month respectively. The second category is unstructured user-submitted intent forms, including content such as user identity information and intended product type.
What constraints these characteristics impose on forms and interactions
Structured product data has fixed sources but updates frequently. This requires that product options in forms be pulled in real time, not written statically, to avoid including options for discontinued products.
Customer intent data is generated in real time on submission. This requires that form submission interfaces support synchronous validation and data storage, to prevent data loss.
Required form fields differ across marketing scenarios. For example, credit marketing requires additional collection of income proof fields, while deposit marketing requires collection of minimum investment amount. This requires that forms support conditional display logic.
When marketing forms are integrated with offline branch traffic sources, support must be provided for automatically filling some known customer information after scanning a QR code, to reduce redundant filling steps.
How to set configurations
| Configuration Item | Recommended Approach | Rationale |
|---|---|---|
dynamic_form_options | Bind to the internal bank product library API, sync once every 12 hours | Adapt to product data update frequency, avoid static options expiring |
form_field_condition | Trigger corresponding required fields based on product type | Match field difference requirements across different marketing scenarios |
UPLOAD_FILE_ALLOW_EXT | ["jpg", "png", "pdf", "docx"] | Cover common qualification certificate and contract file formats for marketing scenarios |
MODEL_API_TIMEOUT | 30 seconds | Adapt to interface response times of internal urban commercial bank systems, avoid timeout errors |
knowledge_retrieve_top_k | Top 3 | Match the precise recommendation needs of urban commercial bank marketing content, avoid excessive redundant results |
form_submit_webhook | Bind to the internal bank customer management system interface | Enable real-time synchronous storage of intent data |
The parameter values provided on this page are common recommended starting points for configuration. Actual values are affected by material form, data volume, and business rules. Specific issues require specific analysis. It is recommended to test on your own samples before finalizing settings.
Three Common Misconfigurations
- The symptom is that multi-level product selection forms fail to dynamically load next-level options based on user selection. The cause is that the linkage logic for
form_field_conditionis not configured, and only a static option list is used. - The symptom is that the system prompts unsupported file format when uploading files via the form. The cause is that
UPLOAD_FILE_ALLOW_EXTis not configured to cover the qualification file formats required for marketing scenarios, and only a small number of general formats are enabled. - The symptom is that variables referenced from the knowledge base cannot be selected in the code run node. The cause is that the output variable binding for knowledge base retrieval is not configured in the preceding node, or the variable exposure switch is not enabled.
How to Verify Proper Configuration
- Open the form preview page, select different product types, and check whether corresponding fields are displayed or hidden according to preset rules.
- Upload test files within and outside the preset configuration range, and check whether the system's upload validation results match the configuration.
- Submit a test form, and check whether the bound webhook interface receives complete intent data.
- Configure a knowledge base retrieval step in the code run node, and check whether the output fields of the knowledge base retrieval appear in the variable selection panel.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-14.