How Alibaba’s AI‑Powered “Future Store” Redefines Unmanned Retail

At the 2022 Yunqi Conference, Alibaba showcased its flagship “Tmall Future Store”, an 80‑square‑meter unmanned retail space that eliminates checkout counters and queues, allowing customers to pick items and leave without any manual payment.

1. Understanding the Unmanned Store Concept

The term “unmanned” does not mean the absence of staff but the automation of repetitive, labor‑intensive tasks through technology. The goal is two‑fold: first, to replace low‑value manual work; second, to digitize the entire offline consumption process, integrating offline and online data for smarter retail.

2. Technical Foundations of the Future Store

The store relies on three core AI capabilities:

Global tracking : real‑time capture of customer movements across the store.

Product recognition : gravity sensors and shelf‑mounted cameras identify items a shopper picks up.

Person‑product matching : linking the identified shopper with the selected products to enable automatic settlement.

These capabilities support features such as automatic checkout, one‑click inventory replenishment, anomaly alerts, supply‑chain forecasting, and personalized in‑store guidance.

3. Architecture – the “6+1” Model

The system follows a “6+1” architecture. Five edge processing nodes (sensors, algorithm engine, execution layer, client devices, and a local gateway) handle data collection, analysis, and actuation on‑site, while the “+1” represents the cloud services that provide transaction processing, data storage, and analytics.

4. Real‑World Deployments

Two case studies illustrate the impact:

Alibaba Library (AliKu) : After conversion to an unmanned format, daily revenue rose 75%, visitor count increased 56.5%, and average exit time dropped to 4.5 seconds for 2,300 daily visitors.

Zhida Bookstore : A 53‑day transformation to version 3.0 boosted daily turnover by 78.3% compared with the previous 2.0 layout.

5. Future Evolution

Alibaba aims to improve three key areas:

Enhance algorithm robustness for varied lighting, angles, and single‑shot face recognition.

Reduce hardware cost by consolidating compute to edge devices and standardizing modules.

Lower deployment time and enable remote, OTA updates via Alibaba Cloud’s LinkEdge platform.

Edge intelligence will allow each camera and screen to perform local inference, minimizing reliance on centralized servers.

6. Vision – Stores as Smart Devices

The long‑term vision is to make every unmanned store behave like a smartphone: continuously updatable, highly interactive, and capable of delivering data‑driven services such as personalized product recommendations, in‑store “埋点” analytics, and seamless online‑offline integration.