Alibaba releases a new AI store assistant, "AI + Human" conversion rate surpasses pure human for the first time

On May 11th, Alibaba released the brand-new AI Shop Assistant, which is the first customer service agent in the e-commerce industry with both pre-sale and post-sale service capabilities. Data from testing shows that after merchants adopt the AI Shop Assistant, the average rate of transferring to human agents decreases by 45%, and the conversion effect of “AI + human” collaboration has increased by over 10% compared to pure human customer service, achieving a breakthrough for the first time. As Tmall’s 618 shopping festival approaches, the AI Shop Assistant officially upgrades customer service to AI-guided shopping, helping merchants grow.

“The customer service industry will officially enter a new chapter of ‘AI-led, human-assisted.’ The AI Shop Assistant is no longer just a ‘question-and-answer machine’ that only responds to inquiries; it now has ‘hands’ and ‘feet’ and can directly convert into productivity,” said Shen Weiye, Vice President of Alibaba Group and President of Taobao Platform.

According to Zhan Lijun, head of customer operations at TaoTian Group, the full-link AI Shop Assistant is built on the latest large model of Qianwen, relying on Taobao’s massive transaction data, and has been fine-tuned for different industries and e-commerce scenarios with multimodal capabilities upgraded. It understands industry rules and user needs better. As of March, the AI Shop Assistant handles nearly ten million conversations daily, with over one million merchants connected.

“AI Shop Assistant + human” conversion effect surpasses pure human customer service

The upgrade of the AI Shop Assistant focuses on three core capability enhancements:

1. It can guide shopping, with “AI + human” conversion rates surpassing pure human for the first time

After the upgrade, the AI Shop Assistant more directly aims to improve inquiry-to-order conversion rates. Under the collaboration of “AI + human,” AI can answer over 80% of questions and proactively offer services to facilitate transactions; for high-value orders that merchants care about most, AI provides decision-making information and order promotion suggestions to human customer service to assist in conversion. In the “AI + human” mode, inquiry-to-order conversion has already surpassed pure human customer service across all time periods, not just at night but throughout the day.

The AI Shop Assistant’s understanding of products and users has also significantly improved, enabling it to proactively anticipate user needs and recommend precisely. It can autonomously call upon tools such as related product recommendations, out-of-stock suggestions, precise payment reminders, and multimodal image recognition to assist in conversion. According to data from the Shop Assistant, after merchants adopted the new version, the average inquiry-to-order conversion rate reached 10%, with apparel merchants seeing an average increase of 20%.

2. Full-scenario management, enhancing store service experience

After the upgrade, the AI Shop Assistant deeply covers over 30 scenarios including product recommendations, size suggestions, national subsidies, and product comparisons. Its scene-based intelligent assistance capabilities have been greatly improved, with response speeds increased by 50% and accuracy rates exceeding 90%. Most inquiries can be automatically answered by AI in seconds, greatly enhancing user experience.

The AI Shop Assistant provides merchants with assured growth support: for example, after Xiaomi’s official flagship store on Tmall adopted it, the rate of transferring to human customer service dropped by 45%, satisfaction increased by 22%; for Xtep’s official flagship store on Tmall, the transfer rate decreased by 55%, and inquiry-to-order conversion increased by 46%.

3. Confident operation, with a successful order recovery rate exceeding 20%

The AI Shop Assistant’s service capabilities have also been significantly enhanced. First, it connects with platform and third-party service provider systems to automatically handle complex processes such as price difference refunds, deposit refunds, product troubleshooting, and address changes; second, it improves risk control and asset protection, such as strictly controlling compensation within merchant-authorized limits to avoid “overcompensation” and “exploitation,” and has integrated the newly launched AI fake image recognition model to accurately intercept “wool party” fraud.

More importantly, when facing potential returns, unlike traditional after-sales handling that is black-and-white, the AI Shop Assistant can accurately identify user emotions and communicate patiently with users like a friend to retain orders. It offers “replaceable solutions” such as size exchanges, style changes, priority troubleshooting, exchanges, or reshipments, reducing returns and refunds, and increasing actual merchant revenue. According to test merchant data, the average success rate of refund recovery exceeds 20%.