As it has in many other Industries, artificial intelligence has become a core component of e-commerce. Going forward, it’s believed that AI and natural language understanding technologies in particular will find many new applications in the industry.
E-commerce is a fast-growing industry. In 2017, global retail e-commerce sales reached US$2.3 trillion, and e-retail revenue is expected to grow to US$4.88 trillion in 2021. The world’s continuous digitalization and informationization gives e-commerce a lot of room for development.
AI technologies are boosting e-commerce sales, enhancing the service experience for users and expanding production scale to create industrial value and new business opportunities. in this Synced report we look at the role being played by natural language understanding and natural language processing technologies in e-commerce.
Related core technologies
Text clustering: Text clustering technology can be used to understand text data in product information content on e-commerce platform webpages. The tech can automatically classify products to improve information and recommendation services.
Machine reading comprehension: Machine reading comprehension can be deployed to understand product descriptions on an e-commerce platform, find core information in the text content, and use the parsed text information and natural language generation technologies (NLG) to compose automated product summaries.
Emotion engine: This technology can analyze consumers’ message content on e-commerce platforms to predict emotions and behaviors. It can encourage positive emotions, and ameliorate negative emotions to improve customer experience.
Named entity identification: This tech makes it possible to quickly find name and entity information of various products by parsing the text content on an e-commerce platform to improve the speed and accuracy of product content information retrieval.
Applications and Scenarios
Some of the applications leveraging NLP in the customer service scenario include: Customer Service Quality Assurance, Detecting Customer Complaint Escalation Potential, Automatically Writing Product Descriptions, and Commodity Management.
Zhongke Huilian Information Technology’s Intelligent Quality Inspection System can evaluate conversations between online customer service systems and customers to measure both service quality and service attitudes. Detecting customer complaint escalation potential is achieved by analyzing customer dialogue content and any expressed emotions. This data is used to judge the possibility of a leap-level complaint, identify users with complaint behavior tendencies, and proactively provide targeted special care to reduce complaint escalation rates.
RSVP.ai has implemented a novel customer behaviour detecting system that combines deep neural networks and traditional feature engineering methods to evaluate customer complaint scenarios.
Weardex meanwhile uses a computer vision system called ViSenze to identify clothing images and help users find their favorite styles online from photo inputs. Weardex is now working on combining ViSense with natural language processing technology to automatically write descriptions of the fashion products from photos.
E-commerce giant Ebay processes and parses data through natural language understanding techniques, and classifies products on its website from different feature perspectives through the application of entity recognition and text clustering. Such classification and identification techniques can make the retrieval of goods more accurate, and at the same time enable more effective product recommendation services. Effective commodity management sees products classified according to different rules, so they can be displayed to customers more effectively. This tech is at the heart of smart product recommendations and smart product search.
Current natural language processing technologies still face limitations in e-commerce, mainly in data bottlenecks. Although a large amount of data is generated in the e-commerce field every day, it is rarely used directly for machine learning model training.
Another bottleneck involves technology. At present, artificial intelligence can only assist humans in writing commodity descriptions, as language models still lack the creativity and passion required for writing compelling commodity advertisements.
In the future, the e-commerce field will continue to generate more and more structured data. Natural language analysis technology will find more applications in e-commerce, bringing more value to the business and more convenience to users.
Author: Ying Shan | Editor: Michael Sarazen