Leveraging computer vision and AI technologies, Comma Smart helps convenience stores reduce labor costs

According to Comma Smart, its system achieved an accuracy rate of 95%, which can save 40% labor costs.

By

Leveraging computer vision and AI technologies, Comma Smart helps convenience stores reduce labor costs

With the rising cost of manpower and land rent, as well as the maturity of commercialization of artificial intelligence and the technology of Internet of Things, unmanned new retail is now a big trend.

In 2017, the industry witnessed the emergence and closing down of a large number of start-ups. Now, with the entry of technology giants, it has come to the stage driven by technology and integrated solution—JD X unmanned supermarket has expanded this year, while Alibaba launched the unmanned shelves “Zhi Xuan” in March. Tencent Cloud also released the unmanned retail solution.

In addition to opening new stores, can new retail stores be rebuilt? Is it difficult to employ the technology? Comma Smart (逗号智能) has already landed and transformed two traditional stores.

At present, unmanned new retail businesses are divided into two models: with or without checkout counters. The technological solutions include RFID, weight induction, and computer vision.

Comma Smart transforms existing convenience stores with its own intellectual properties: artificial intelligence + CV technology. The transformation can be completed within 1 day—it will provide hardware (could be based on the original store equipment) + software (its self-developed plug-in enables user data to be connected with China’s first 5 ERP systems and SaaS system) to improve the efficiency of the traditional convenience stores and reduce the human cost.

According to Comma Smart’s founder and CEO FENG Xinyu, after more than three months of testing by tens of thousands of people, its complete set of system achieved an accuracy rate of 95%, which can save 40% human cost for the store. The cost of the whole package of the solution is tens of thousands of yuan, including:

  • Intelligent access control system: with face recognition door opening function, it can recognize hundreds of thousands of human faces within 1 second at a time.
  • Intelligent payment system: image recognition technology enables the system to identify more than 1000 SKUs with less than 0.3 seconds of recognition time and the accuracy rate of up to 95%.
  • Intelligent checkout counter: three cameras can realize multi-angle check and can scan 5 pieces of goods at the same time; the solution will support face authentication of entity’s biology information and will allow the AI-based “hand gesture payment,” enabling members to face the lens and give a thumbs up rather than scan QR Code to complete the payment.
Image credit to Comma Smart.

According to FENG, although Comma Smart can realize completely unmanned operation with current technology, it suggests the stores should employ a small number of shop assistants and use its solution as an aid to deal with complicated and flexible missions and improve efficiency.

At present, Comma Smart’s revenue is mainly from the sale of the whole solution. In the future, the hardware is expected to be the entry point of market and data. Payment for purchase is not required, but the startup will charge for the deposit and monthly service fees of SaaS, which will be used to connect the data of small and medium convenience stores.

FENG Xinyu introduced that the long-run profit plan is to digitize the sale and consumption of the convenience store in order to perform accurate advertising, guide the merchant distribution as well as marketing and production of supply chain.

According to Comma Smart, the company now enjoys the capacity to realize large-scale replication. Other than the current two demonstration stores, a smart store in cooperation with Zhouheiya, a Chinese food company, will open at the end of this month. In 2018, 2,000—5,000 offline stores are expected to be transformed.

Currently, the team consists of nearly 100 people. The co-founder and CEO FENG Xinyu, who is a serial entrepreneur, has 17 years of Internet-related experience.

Writer: MO Li