Intelligent outbound robot best practices
Workflow
1. Enterprise qualification
1.1 Enterprise real-name authentication
After you register your account, follow the instructions in the Enterprise Real-name Authentication guide to complete the enterprise real-name authentication. You will need the following documents:
Business License Duplicate
Legal Representative ID Certificate
Corporate Bank Account Verification
2. Product activation
Go to the Place Order page to place an order and activate the service. For pricing details, see Product Billing.

After the service is activated, you can access the Intelligent Outbound Robot console. When prompted for STS authorization, click Authorize.
3. Permission configuration
3.1 RAM user authorization
Follow the instructions in Create and authorize a RAM account to create a RAM user for operations:
Create a RAM user in the RAM console.
Attach the OutboundBot product access policy.
Apply the principle of least privilege. We recommend using separate permissions for development, testing, and production environments.
NoteWe recommend using a RAM role to isolate permissions. The main account should retain full administrative privileges.
4. Basic space creation
4.1 Business space creation
Create a Conversation Engine 3.0 space in the Business Management section:
Notes:
The legacy
chatbotoption only supports a non-commercial version without large model capabilities and has been deprecated. Always select Conversation Engine 3.0 to build your space.We recommend allocating
instance concurrencybased on the call rate limit of your numbers. For example, if a number is bound to a single business space and has a rate limit of 10, allocate aninstance concurrencyof 10 to that space. If the same number is bound to two business spaces, you must divide the concurrency. For example, allocate aninstance concurrencyof 5 to each space.
5. Bot scenario building
5.1 Scenario types and configuration
Type | Scenario | Mode | Guide | Use case |
Notification Interaction | One-time voice playback or DTMF input | Small Model Scenario Management |
| |
Intelligent Conversation | Multi-turn human-bot conversations | Text Mode |
| |
Complex Flow | Flows requiring complex, stateful logic | workflow configuration mode |
| |
External Model | Integrate with third-party large models | Function Compute Mode |
|
Note:
When building scenarios that use large models, if your flow requires variables, number collection, an API plugin, or the option to transfer to a human agent, you must first define them in the Variable Configuration section.
5.2 Live demo
Experience a live demo at Lingque Scenario Examples.
6. Testing and validation
6.1 Comprehensive testing
Perform voice debugging and testing:
Use the voice debugging tool to test TTS and ASR performance.
Use the demo link to validate the scenario.
7. Number resource management
7.1 Number application and binding
To apply for numbers, see the Line Application Process documentation.
When you apply for a phone number:
Allow 1 to 2 weeks for review and number provisioning.
You must provide a clear description of your use case.
Bind numbers to a business space:
A single number can be bound to multiple business spaces.
We recommend creating a number pool and grouping numbers by business type.
7.2 Concurrency configuration
Allocate concurrency in the Business Management section based on your business needs:
The
instance concurrencyfor a single business space must not exceed the combined call rate limit of all its bound caller ID numbers.
8. Task execution management
8.1 Task group creation
Bind a published scenario template and a caller ID.
Set a call strategy for the task. The calling schedule must comply with the permitted calling hours of your phone line.
Upload a list of called numbers.
9. Quick integration
For details on integration solutions, using APIs to invoke tasks, and receiving callbacks with task results, see Quick Integration.