Yili Group has deployed AI agents across its front-line sales operations-shopping guides, community management, and influencer marketing-using Tencent Cloud's Agent Development Platform (ADP). The rollout supports tens of thousands of shopping guides who field a heavy volume of specific, real-time consumer questions every day. Early data shows a 15.7% increase in product link click-through rates in shopping guide communities and a 26% rise in shopping guide orders.
AI moves into high-touch sales roles
The shopping guide scenario is the most direct. In brand communities, a single guide faces questions about ingredients, promotions, and purchase timing. Before, she had to hunt down product information and promotion rules herself, then craft a response from experience. Yili built a shopping guide agent that pulls together product knowledge, promotion policies, and domain expertise-like an always-online expert beside the guide.
Wu Chunze, general manager of Tencent Smart Retail's vertical industries, said "today's consumers are armed with AI and becoming more and more professional." A mother asking about milk powder for a sensitive stomach may have already read ten reviews and compared three ingredient lists. When she asks the guide, she is verifying the brand's professionalism. The AI agent helps the guide answer with speed and depth that was hard to sustain at scale.
Influencer marketing saw a similar shift. Business staff previously had to find influencers one by one, message them, discuss terms, and schedule posts-managing two or three per day at most. Yili's influencer marketing agent now handles initial outreach, sharing cooperation policies, and standardized follow-up. Only influencers with clear intent get passed to a human for deep discussion, freeing staff for higher-value work.
Training the agent for the front line
Shang Zhihu, general manager of Yili Group's Digital Technology Center, described the approach as "consumer-centric and business-oriented." The company started its AI push in 2021, targeting AIGC content generation, vibe coding, and intelligent customer service. The real work, though, was preparing the agent for the specifics of Yili's business.
First, the knowledge base had to be organized. Product formulas, target groups, promotion policies, membership benefits, and compliance language were scattered across systems and inside the heads of top performers. Both sides spent significant time standardizing information so the AI could accurately retrieve and use it. Second, boundaries were defined. Shang Zhihu said the agent "mostly provides strategies and serves as an assistant. In important internal business processes … humans should make the final decisions."
Wu Chunze compared the agent to "a newly recruited digital employee" that needs clear harnesses and guidance. The Tencent Cloud platform acts as a factory for AI Agents & Automation, letting Yili define agents, configure rules, and run pilots quickly while Tencent handles security, compliance, and data backup. Because most interactions happen inside WeChat's ecosystem-enterprise WeChat and mini-programs-the agent fits naturally into the workflow.
Measurable impact on sales performance
The numbers are starting to tell the story. Product link click-through rates in shopping guide communities climbed 15.7%, and shopping guide orders rose 26%. For a newly hired guide in the maternal and child category, who typically needs three to six months to learn formulas, allergens, and feeding methods, the agent shortens the ramp-up. She can identify needs faster and give professional responses sooner.
Wu Chunze noted the effect on trust: "The consumer feels that when chatting with you, you understand him well and are very professional. The shopping guide has become more professional and caring because of AI, which has also brought in more sales and higher income for herself." The integration of AI for Sales here is not about replacing staff. It handles repetitive, knowledge-intensive tasks so humans can focus on judgment and relationship-building.
Why this matters for sales professionals
Yili's experience points to a practical playbook. First, identify the high-frequency, knowledge-heavy interactions that slow your team down-questions about specs, pricing, inventory, or compliance. Second, invest time in organizing the knowledge that already exists in your systems and your best people's heads. Without clean, standardized information, an AI agent will give unreliable answers. Third, define clear boundaries: let the agent provide options and information, but keep decisions and sensitive judgment with humans.
The result is a sales force that responds faster, sounds more expert, and builds trust at scale. As Shang Zhihu put it, "Suppose a person's productivity was 100 before. After using these agents, it might become 2000." For sales teams, that means more time for the conversations that actually close deals.
Your membership also unlocks:
