How RPA Powers 24/7 WeChat Smart Customer Service for E‑commerce Ops

Background

Rapid business expansion has led to a growing number of merchants, and the after‑sales team communicates with customers primarily through enterprise WeChat groups. Limited team resources cause repeated questions, delayed responses during nights or weekends, and a poor user experience, especially during peak promotional periods.

Challenges and Considerations

JD.com’s internal knowledge assistant cannot be applied to the WeChat side.

Enterprise WeChat does not easily integrate third‑party knowledge bases; image‑based messages are hard to process.

The built‑in WeChat chatbot does not support enterprise groups containing WeChat accounts, and a single knowledge base per enterprise becomes chaotic when many business lines coexist.

Need to enable merchants to @ the service account for targeted replies while monitoring over 200 groups in real time.

Require dynamic configuration to control monitored groups and to forward unanswered queries to human agents.

Maintain merchants’ habit of using WeChat for communication without forcing a new platform.

RPA Overview

The automation robot can be programmed with command scripts to simulate human actions, enabling full desktop application automation—including ERP, browsers, CRM, WeChat, DingTalk, and any other daily tools. It can automate web tasks such as JavaScript execution, data extraction, form filling, and API calls, while controlling keyboard and mouse just like a person.

Implementation Steps

Build a knowledge assistant that continuously collects and refines merchant questions.

Gather similar issues and test coverage of the training set.

Visualize answers by invoking the knowledge base, distinguishing text from images, and normalizing elements for accurate RPA capture.

Deploy RPA to monitor WeChat groups 24/7, with dynamic monitoring parameters.

Capture @‑service messages in real time.

Provide a Q&A interface for agents to query and copy answers.

Use Excel for dynamic configuration of monitored groups and to log recent chats for deduplication.

Connecting WeChat and the Knowledge Base

RPA extracts elements from the Q&A interface, parses answer content, and separates text from images.

RPA copies the answer (text and images) into the WeChat group, sending a consolidated reply.

Persist group questions in Excel.

Read Excel to obtain dynamically configured group information and human‑agent contacts.

Results

The WeChat smart customer service now monitors dozens of groups, overcoming WeChat’s native limitations, delivering 24/7 automated operations, and significantly easing the after‑sales workload. During major sales events, the system scales to monitor additional merchant groups, ensuring stable, high‑quality service throughout promotions.