How to blend Artificial Intelligence with Live Commerce

Customer experience is the key factor for a business to emerge from increasing competition. Though 84% of executives agree on the importance of customer experience and prioritize customer-centric strategies, only 14% of them report that they have strong capabilities in that area. Actually, today’s current business context has imposed on companies a series of strategies and adaptations that have become mandatory in order to survive the continuous market stress: the customers change their opinion, they’re really attentive, knowledgeable, and conscious so companies have to be able to understand their needs in a timely manner.
As a matter of fact, to minimize bad user experiences, increase profitability, and abandon clients’ dissatisfaction, businesses have to know their customers, and the only way to do so is to adopt Customer Intelligence.

Customer Intelligence can be defined as the ability of a company to collect customer data from heterogeneous sources ( email, call-center, social network, forum, apps, chats, etc.) to draw a “picture” of the client in all its facets, thus optimizing sales&marketing and CRM strategies, and improving business performance through targeted and strategically successful decisions. Therefore, the customer intelligence process contains a collection of customer data via multiple channels and includes the use of technologies such as feedback management, social media monitoring, Natural Language Processing (NLP) as well as other advanced analytics and data management technologies.

To provide sophisticated Customer Intelligence, we’ve integrated NLP with shoppable videos and contents. Indeed, GoLive is the first end-to-end solution capable of combining the innovative power of Live Commerce — an original online shopping feature that allows retailers to sell their products during a live stream — with real-time analytics and artificial intelligence: killer feature for which GoLive stands out from the crowd. The goal is to enable businesses to analyze customer interactions and provide actionable insights and predictive analytics in real-time to enhance the customer experience related to a new digital shoppertainment format.

GoLive leverages an analytical process that uses real-time analytics and artificial intelligence algorithms to extract, at each event, insights always updated and accurate. The platform collects, analyzes, and sorts chat anonymized data into relevant categories bringing factual benefits to companies, allowing them to understand and predict in real-time what customers are actually looking for, want or what they are thinking about, with a significant impact on the entire value chain.

Specifically, our work provides higher-level NLP capabilities, such as sentiment analysis, emotional analysis and topic discovery and modeling. Let’s see in deep what they mean.

Sentiment Analysis

The sentiment analysis measures the attitude of the customer towards the aspects of a service or product which they describe in text. This typically involves taking a piece of text or comment and returning a score that measures how positive or negative the text is.
By using sentiment analysis during a live show, a Brand can gauge how customers feel about different aspects of its products without having to read thousands of customer comments at once. Our approach to sentiment analysis consists of recognizing the sentiment based on the words and their order using a sentiment-labeled training set and a deep learning technique.

Emotional analysis

Contrary to sentiment analysis, emotional analysis relies on a more sophisticated and complex system. While the first one uses a simplified categorization (with two or three categories), the latter relies on a deeper analysis of human emotions and sensitivities. This method highlights the nuances between the different feelings readers express. Inside positive it detects specific emotions like happiness, joy, or excitement. While sentiment analysis helps a Brand to know how its content is performing, the emotional analysis leads to understanding why. In GoLive, the emotional analysis is conducted by analyzing emojis and emoticons used in the chats, and 9 different types of emotions are extracted.

Topic discovery and modeling is conducted to accurately capture the meaning and themes of the questions in text collections. In GoLive topic modeling is conducted through an unsupervised machine learning method to all the questions made by the viewers during a live show. The questions are then grouped into 7 different themes like fabrics compositions, live show information, clothes sizes, etc.