How retailers benefit from conversational commerce and generative AI

Man with smartphone chats with AI chatbot

Selling responsively: The use of chatbots and AI in
e-commerce

It took just two months for the AI-based chatbot ChatGPT to crack the 100 million monthly active users mark after its launch. The program demonstrates what artificial intelligence (AI) is capable of. Generative AI, as used by ChatGPT, can create high-quality content – and do so in a very short time. Especially in e-commerce, the potential of generative AI is enormous. As an expert in conversational commerce, OSP supports companies in the implementation of generative AI. Learn about the added value of generative AI based on two current projects.

Good to know: Generative AI refers to a class of artificial intelligence systems that are trained to generate previously non-existent data based on data with which they are trained. To do this, generative AI uses techniques such as neural networks to recognize patterns in data and generate new content.

More than just a prop: Generative AI as a means of customer retention

 

Generative AI cannot replace humans. It is a practical tool that helps companies simplify and accelerate processes. It is also capable of relieving employees in one area in particular: conversational commerce, i.e. communication between companies and their customers – and around the clock.

Before we present two OSP use cases that illustrate the potential of generative AI in conversational commerce, we first take a general look at the topic of conversational commerce and summarize what lies behind it. 

 

What is conversational commerce?

 

The term conversational commerce refers to communication between companies and their clientele on channels such as messaging apps, chatbots, social media, and voice assistant technologies.

The goal of conversational commerce is to accompany and support customers along their customer journey – from the discovery of products or services to the purchase decision.

Conversational commerce has become an important part of modern e-commerce strategy, as it offers the opportunity to communicate efficiently and directly with the clientele online, which in turn promotes customer satisfaction and loyalty.

 

Usage of conversational commerce


Here are some of the key application areas of conversational commerce:

  • Customer service:
    The main area of application for conversational commerce is customer service. Chatbots and automated messaging systems provide real-time answers around the clock.
  • Personalization:
    Through direct communication with the clientele, the system understands the preferences and needs of the users and proposes personalized offers and solutions.
  • Sales promotion:
    With the help of the mentioned personalized product recommendations, users feel directly addressed and the likelihood of making a purchase increases.
  • Transactions:
    Another strength of conversational commerce is the ability to conduct transactions directly within a messaging channel. This makes the buying process seamless and easy for users.
  • Customer loyalty:
    Real-time communication through messaging channels helps build strong customer loyalty.
  • Scalability:
    With the help of automation and AI, companies can serve thousands of customers at the same time without compromising the quality of service - for example, during a live shopping event.

 

Some well-known platforms used for conversational commerce are WhatsApp, Facebook Messenger, Telegram, WeChat and many other messaging apps, as well as voice assistants such as Amazon Alexa or Google Assistant.

 

Generative AI in use at OSP: Our use cases

 

Whether communication via WhatsApp, Messenger, or other messaging services: many e-commerce companies are already relying on the advantages of AI in the form of chatbots. But how can the chatbot concept be further developed in the context of conversational commerce and combined with modern variants of e-commerce? OSP proceeded this route and developed two useful chatbots with real added value for companies.

 

OSP use case 1: AI Chatbot in live shopping

 

As a pioneer and professional in the field of live shopping, the IT company OSP developed a chatbot that supports presenters of live shopping events. The bot provides answers to possible questions from viewers and makes it easier for the company to respond to requests during the stream. If necessary, the moderators can additionally adapt the suggestions of the messenger.

Technology at its best: The OSP bot used in Live Shopping is based on Microsoft Azure, Cognitive Services and ChatGPT and was programmed using the GitHub Copilot.

The chat bot, which was integrated into OSP's own video commerce platform MOVEX | Live Shopping, was tested in the live shows of the client OTTO. Already in the first test rounds, the program delivered high-quality answers in the interaction with users. This was made technically possible by prompt engineering and an Azure OpenAI instance.

At OTTO, the AI feature is already integrated in the moderation area of the live shopping shows.  In addition to the live shows, the chatbot can also be used in the video-on-demand area. Here, the chatbot for conversational commerce will provide advice around the clock and answer questions about the respective shows.

 

OSP use case 2: Generative AI helps with e-commerce shopping

 

Another bot developed by OSP is also available around the clock. It can be integrated in various contexts, for example in the web stores of companies and brands. The program in the form of a chatbot recommends personalized products to users and helps them with questions about orders, shipping and returns.

The focus here is on interaction with the customer – very much in the spirit of conversational commerce. However, the bot goes one step further: The AI recognizes the intention of the user: Do they want a recommendation? Are they looking for information about a product? Or do they want to find a comparable item to the one displayed?

Matching the user's intention, the AI searches and filters the assortment, compares data on products with information from the conversation with the users, and finally provides title, link, and images of eligible articles. During the consultation, the bot takes into account the entire conversation and the entire range of the respective retailer.

More than just a conventional chatbot: Compared to the chatbot for live shopping, a model is used here that consists of several stages of interaction. In addition to Python/Flash and Azure OpenAI, the program uses various models such as GPT-3.5 (for creating descriptions) or Ada Embedding to recognize matching items using semantic vectors.

The current prototype of the Messenger bot is already running and collecting data via crawler, but development is still ongoing: Soon, the chatbot should recognize prompts even more reliably, pull data from the assortment more efficiently, and locate items within the shortest possible time.

 

Chatbots are just the beginning in e-commerce 

 

The two conversational commerce solutions developed by OSP show the potential and possibilities of chatbots and co. in e-commerce. Almost one in four companies in the retail sector is already using AI in 2023 (by comparison, the figure was 7.5 percent in 2020), and the trend is rising.  Software used in the right places helps companies improve support and service with the help of messengers and save costs.

To ensure that customer trust and satisfaction do not suffer, it is important to use technology responsibly. If companies take this aspect into account, not only can the concept of Conversational Commerce: AI have the potential to revolutionize e-commerce.

Become part of the technological revolution:

OSP supports you in word and deed to profit from the advantages of generative AI. Let us advise you without obligation – or inform yourself in advance: In our free whitepaper "Artificial Intelligence in Digital Commerce" you will learn everything about the opportunities of AI for online commerce.

Please note the disclaimer regarding the use of the AI assistant.
text bubble How can I help you?