Use Alibaba Cloud Model Studio and Chat App Message Service to build a WhatsApp intelligent reply bot. This bot uses your enterprise knowledge base to provide 24/7 intelligent customer service and shopping guidance. This tutorial guides you through the entire no-code process, where you will upload a PDF or Excel knowledge base, create a retrieval-augmented generation (RAG) application, and connect it to your official WhatsApp account.
Related policy
Starting January 15, 2026, Meta will ban the deployment of General-Purpose AI Chatbots on the WhatsApp Business Platform (API). This new rule applies to new API users starting October 15, 2025. For more information about the policy, see WhatsApp 2026 AI Chatbot Policy Explained | Alibaba Cloud Chat App Compliance Guide.
Preparations
Prepare your enterprise knowledge base files on your local machine. These files serve as the knowledge source for the large language model (LLM) application and can include unstructured data, such as DOC, PDF, MD, or TXT files, or structured data, such as XLSX or XLS files.
Enterprise knowledge base files:
File type | Description | Example |
Unstructured data file | Document-based knowledge files suitable for systematic knowledge input. These are complete documents that contain systematic knowledge, such as product manuals, technical documents, or user guides. The system processes these files to extract knowledge. Users can then query for details from the documents. You can upload DOC, PDF, MD, and TXT files. |
|
Structured data file | Q&A-style knowledge files suitable for providing specific answers. These files contain knowledge points as question-and-answer (Q&A) pairs. This format is ideal for directly answering specific questions. It is concise and targeted, which helps users get answers quickly without reading a full document. The system can use these Q&A pairs directly to respond to user queries. You can upload XLSX or XLS files. |
|
Procedure
Step 1: Build a private Q&A application
For more information, see Build a no-code Q&A application.
Replace the prompt from Step 1 with the following:
You are an expert in handling ChatFlow product inquiries. You specialize in answering customer questions about the product's features. Do not answer questions about other topics.
Provide concise answers based on the knowledge base. Each response must not exceed 600 words. Use emojis as needed to make the tone more friendly.
Step 2: Configure the flow in Chat App Message Service or the CPaaS platform
Prerequisites:
You have completed Alibaba Cloud account registration and enterprise verification, and enabled the Chat App Message Service.
You have created a WhatsApp channel.
You have registered a WABA. The WABA is approved by WhatsApp and the phone number is connected.
The CPaaS platform is ideal for service providers because it simplifies management and distribution. For more information, see CPaaS Platform.
Chat App Message Service
You can use the flow editor in Chat App Message Service to enable multi-turn intelligent conversations with the LLM.
Log on to the Chat App Message Service console. On the Flow Editor page, under View Templates, select Multi-turn conversation using a large language model (LLM).
CPaaS platform configuration
You can use the flow editor in the CPaaS platform to transfer conversations between the automated intelligent reply system and human agents.
Log on to the CPaaS console. On the Flow Editor page, under View Templates, select Customer service with a large language model (LLM) and human agents. This template enables transfers between the Chat Flow and human agents.
The following steps are the same for Chat App Message Service and CPaaS:
On the flow editor canvas, find the Natural Language Generation component.
In the Natural Language Generation configuration panel on the right, enter the API-KEY and Application ID (`appId`) that you obtained. Set Application Type to Application. If you are using Alibaba Cloud Model Studio on the China site, set Protocol to Dashscope. If you are using Alibaba Cloud Model Studio on the International site, set Protocol to Dashscope(International). Then, click the Create Flow button. The WhatsApp AI assistant chat flow is created.
Step 3: Publish the Chat Flow
After you create the Chat Flow, click Save and Publish. Then, click Bind and Publish.
On the WABA Binding page, select your WhatsApp Business Account and phone number.
Your WhatsApp AI assistant is now ready to serve users.
