news

"to fully tap business opportunities, the retail industry needs to analyze a large amount of data, which requires us to be able to collect the data, save it, and access it when needed. therefore, the industry vertical based on the general large model base the creation of large models is particularly important." recently, during the 2024 amazon cloud technology consumer technology summit, zhang xiangyang, general manager of shenpan technology of new hope fresh life group, said in an interview with media including the "daily economic news".

according to zhang xiangyang, in the era of information and digitalization, companies have been exploring how to find optimal solutions to business problems, which usually involves theoretical support from operations research, quantitative economics, and statistics. however, as the amount of data and considerations increase, traditional methods gradually expose problems of insufficient efficiency and computational bottlenecks. therefore, new generation computing methods such as intelligent learning have emerged as new ways to solve complex retail problems.

in recent years, new technologies are bringing about changes in the consumer retail scene: the introduction of digital technology, artificial intelligence, especially generative ai technology is gradually reshaping the traditional "people, goods, and place" relationship, pushing the industry to become more intelligent and efficient. and humane development.

in this context, cloud vendors represented by amazon cloud technology, including bytedance's volcano engine, and jd cloud, are actively promoting the application of ai in retail scenarios. they are all trying to digitize the retail industry. find room for business growth.

amazon cloud technology lays out the retail industry

amazon cloud technology technicians told reporters that in traditional rule-based and mathematical methods, when the number of variables increases, it is difficult to perform effective fitting through mathematical methods. ai technology, especially the application of large models, can handle more variables and data, and build models that can cope with complex data environments through automatic feature extraction and model parameter tuning. in order to achieve further implementation in consumption scenarios.

in fact, the cooperation model between amazon cloud technology and canpan technology is particularly typical at present. the underlying logic of the cooperation between the two parties is that amazon cloud technology provides the underlying general large model capabilities, while retail service providers such as canpan technology build on this basis. create large industry models belonging to specific scenarios to call and support underlying capabilities such as line scheduling, warehouse inventory, driver status, line adjustment, and temperature monitoring involved in the entire supply chain and logistics.

during the interview, the reporter learned that the cooperation between shenpan technology and amazon cloud technology mainly focuses on the three core areas of "warehouse", "distribution" and "people". "warehouse" emphasizes inventory turnover and improves product circulation efficiency; "distribution" emphasizes intelligent transportation routing and fast delivery; "people" emphasizes improving the work efficiency of employees.

these three major areas in the retail industry have a unified underlying element - data.

zhang xiangyang told reporters that in the early days of the information age, data maintenance work in the retail industry mainly relied on manual work. however, this traditional method has two major drawbacks: first, it is inefficient and can often only achieve post-event management; second, data coverage is limited and mainly focuses on necessary business data, such as distribution in traditional erp (enterprise resource planning) systems. layer data. with the rapid development of iot technology, this situation is changing.

zhang xiangyang, general manager of shenpan technology. photo source: photo by intern reporter zhang zitong.

nowadays, enterprises rely less and less on manual data maintenance, and more and more use iot devices to directly collect data. no matter which business link you are in, the required data, such as temperature, number of door openings, document information, etc., can be collected in real time through the corresponding iot devices. this transformation not only greatly improves the efficiency and accuracy of data collection, but also makes data coverage broader and deeper.

taking the logistics industry as an example, the traditional cargo handover and inventory process requires a lot of human participation, but now, these links have been basically unmanned. through video equipment and iot sensors, the handover and inventory process of goods can be completed automatically, and data is uploaded to the system in real time, greatly improving work efficiency and accuracy.

"the introduction of internet of things technology allows us to understand the actual conditions of hundreds of warehouses across the country in real time, including the operation and temperature of online vehicles. this is a huge improvement for us, which not only improves work efficiency, but also makes our decisions are more accurate and timely," zhang xiangyang said.

taking fresh products as an example, zhang xiangyang said that in order to achieve extremely fast delivery of fresh products, it is necessary to immediately arrange intelligent wiring, full-link real-time temperature monitoring and early warning, etc. this requires manufacturers to collect data in real time and provide timely deal with.

but at the same time, zhang xiangyang also bluntly said that the two parties also faced some challenges during the cooperation process. for example, during a warehouse inventory, cameras need to be installed to collect data, but there is a conflict between the warehouse's electrical wiring specifications and the camera installation requirements. faced with this dilemma, the two parties finally found a solution that not only complied with regulations but also improved business through full communication and negotiation.

cloud vendors compete for retail scenarios

as the scale of the global cloud computing market continues to grow, digital transformation in retail consumption scenarios will become more profound. vendors such as volcano engine, jd cloud, alibaba cloud and amazon cloud are also actively promoting the application of ai in retail scenarios.

in august this year, volcano engine released a series of product upgrades for bean bag models, and joined hands with a number of retail companies to establish the "retail model ecological alliance". zhang wenzhong, founder of multipoint dmall and founder of wumart group, said bluntly at the time that the retail large model ecological alliance is a way for retail companies to stay together for warmth. "we must fully embrace ai, not just for a better future, but for survival."

tan dai, president of volcano engine, previously said in an interview with a reporter from "daily economic news": "this time i chose retail because i feel that retail covers enough people and data, and we believe that big changes can be made here. , and there are huge industry barriers and know-how (technical know-how) for terminals, which is also worthy of innovation.”

jd cloud relies on its traditional supply chain advantages to provide merchants with a full range of digital intelligence solutions in retail consumption scenarios.

recently, jd cloud proposed ten major industrial scenarios, including digital smart retail. through the release of five major industrial platforms, including supply chain control tower, energy carbon cloud, and automotive supply chain collaboration platform, jd cloud has further consolidated the product matrix of the digital intelligence supply chain and enriched the industry segmentation scenarios of the digital intelligence supply chain. in the retail field, jd.com’s yunyanxi digital human live broadcast has become a highlight this year.

obviously, major cloud service providers are trying to bring unprecedented changes to the retail industry through ai, big data analysis and other means. it is foreseeable that the retail scenario will become another important battlefield for cloud service vendors to realize business value. however, in this competition, we cannot just look at the advanced level of the general large model base. how to deeply understand the retail business process and provide customized solutions that truly fit the characteristics of the industry is the real focus of this competition.

daily economic news