Tencent Cloud Agent Development Platform (Tencent Cloud ADP) Workflow refers to the important information extracted from dialogue as parameters. Parameters can be nouns, place names, phone numbers, hospital names, etc.
The parameter extractor node belongs to the Information Collection Node. It collects parameters through multi-round dialogue with the user. Parameter information must be defined based on your business scenario. Workflow execution will be interrupted here for interaction with the user until all required parameters are collected.
Directions
Input Variables
Input variables take effect only within the same node and cannot be used cross-node. Support up to 50 input variables to meet multi-input scene requirements. Click "Add" and configure as follows to add input variables.
Configuration
Description
Variable Name
The variable name can only contain letters, digits, or underscores, must start with a letter or underscore, and is mandatory.
Description
The description of this variable. It's optional.
Data source
The data source of this variable supports two options: "refer" and "input". "Refer" allows selecting output variables from all preceding nodes, while "input" enables manually filling in a fixed value.
Type
The data type of this variable is unselectable and defaults to the variable type of "refer" or the string type of "input".
Note:
The parameter extraction node uses three system variables—"SYS.UserQuery", "dialogue historySYS.ChatHistory", and "SYS.CurrentTime"—to perform parameter extraction. Therefore, the system defaults to importing these three variables with no need for manual addition.
Collect Parameter
Model
You can set the LLM used by the node to extract required parameters from user dialogue, supporting only selection of LLMs with permissions for that account.
Parameter Information
Click "Add" and configure as follows to extract the required parameter information.
Configuration
Description
Name
The parameter name, recommend using a specific name, making it easy for model recognition and understanding. Mandatory.
Parameter Type
The data type of the parameter. Mandatory.
Parameter Description
The description of the parameter helps the model accurately identify and understand it. It is advisable to fill in as "concept definition + value requirement". You can use AI for one-click optimization to adjust the content. Mandatory.
Parameter collection example
The correct and incorrect examples for this parameter. The correct example will be used to prompt the LLM on which parameter values will be extracted, while the incorrect example indicates which parameter values must not be extracted. Optional.
Required
Is this parameter required. If the parameter is required, then when the user dialogue does not provide the parameter information, the system will automatically generate a counter-question to ask the user; if the parameter is optional, the system will not perform a counter-question when the user does not provide the parameter information.
Prompt
If you have special requirements for the parameter extraction process, you can configure the prompt in "collect parameter" to guide the LLM to generate appropriate counter-question clarification responses. Typical scenarios include:
1. If no parameter is extracted from the dialogue, inform the model how to counter-question the user.
Prompt example: Follow-up question response when the "registration ID" parameter is missing:
If the "registration ID" parameter value cannot be extracted, please reply with the following content to request the parameter value: "What is your registration ID? You can find the registration ID on your appointment slip. Please provide this information accurately, otherwise the registration process cannot be completed."
2. Reply content has specific format requirements or preferences
Prompt example: Response for the "reservation user's name" parameter:
If the "reservation user's name" parameter value is extracted and the value is a single surname, please reply to the user in the following format: "Mr./Ms. xx". For example, if the value is "Li" or "Surname Li", politely address the recipient as "Mr./Ms. Li".
Output Variable
The output variable processed by this node defaults to the configured parameters for user, as well as runtime Error information (data type is object, this field is empty during normal operation). Manually add is not supported.
Application Example
When users need an invoice, use the parameter extraction node to collect required order number, membership card ID, and username.
Parameter Extractor Node configuration as follows:
FAQs
If you wish to copy configuration information of other parameters, you can click the "Copy" icon in the upper-right corner of the new configuration parameters box and select the parameters you need to copy.
