Introducing the latest developments of LLMs (Large Language Models) in E-Commerce in 2024. There is intense focus on improving the shopping experience with AI, such as Amazon's Rufus, Etsy's gift mode, and Shopify's Sidekick. We’ll also introduce Allganize’s recommended LLM which, unlike others, does not require separate model training.
Amazon has been steadily rolling out features that bring AI to shopping. Last January, they launched four features to help customers buy clothes that fit their bodies well and a feature that bundles common content from product reviews.
Amazon's recently launched Rufus is an AI-based shopping assistant that learns from product manuals, customer reviews, community Q&A, and information on the web. By chatting with Rufus, you can search for products, compare products, and get recommendations on what to buy. Follow-up questions can help you narrow down your search or ask more specific questions.
Amazon is said to have built their new LLM specifically for E-Commerce, training it on its own data as well as data published on the web.
Here are some things you can do to help you with Rufus:
Amazon Rufus plans to improve its AI model through customer feedback. You can rate answers with likes/dislikes and even provide free-form feedback. We are excited to see what impact Amazon’s shopping-specialized LLM will have on improving customer experience and increasing sales.
Etsy launched “Gift Mode” in January, reflecting its collection of over 200 gift recipient personas and 100 million products. You can choose who you are shopping for (parents, children, siblings, etc.), what the purpose is (birthdays, anniversaries, etc.), and what the recipient's interests are. It is organized around over 200 different personas, including maximalists, dog family members, and music lovers. It is said that personas are continuously added to keep up with new trends.
Gift Mode was created using a combination of machine learning, human curation, and OpenAI Chat GPT-4 . GPT-4 helped in configuring various gift purposes and personas, and the gift mode was expanded based on Etsy's existing machine learning and inference stack. In the case of gifts, the results were adjusted to ensure that sellers providing excellent customer service and high-quality product lists were shown first.
Etsy has been investing in strengthening its AI search function for a long time, and also acquired the natural language processing startup Blackbird Technology . Because Etsy has a large number of small, diverse sellers (more than 6 million people), they have focused on giving customers a better browsing experience across thousands of product listings.
Etsy CEO Josh Silverman announced in his May 23 earnings call that he would change Etsy's user experience through an LLM. It remains to be seen how well the Gift Mode created by Chat GPT-4 will be received by Etsy's 90 million customers.
Many companies, including Meta, Google, and Amazon, have allowed advertisers to use generative AI to generate backgrounds for product images since last year. Shopify launched Magic Media Editor in February, which helps sellers edit the background and scenes of product images with generative AI. You can choose premade backgrounds from seven styles - Minimal, Vibrant, Natural, Urban, Rugged, Refined, and Surreal - or by create something entirely new by entering a prompt.
In June 2023, Shopify launched Sidekick Bot, an LLM tool for sellers. A conversational AI assistant trained to understand everything about Shopify, It can answer questions about why sales are down, generate a report summarizing your top 10 best-selling products, or guide you through setting up effective email campaigns to send to your customers.
The LLM-based " Shopify Magic " is said to have been created by combining Shopify's proprietary data, including seller business data, with a publicly available LLM service.
So far, a core set of functionalities and workflows related to LLM-based content creation and operations are already included. With focus on AI functions for sellers, such as creating product description pages suitable for SEO (search engine optimization), automatically generating FAQs and responses, and converting real-time chat with customers directly into payments, Shopify is providing a valuable service to its constituents.
For B2B sellers, the company is expected to launch an automated store that can use the Shopify API for back-end access and various customized front-end solutions. The overall strategy is transforming Shopify into a technology company that increases sellers' profits and keeps them on its platform.
The AI revolution in shopping, however, does not stop at online retailers. An AI assistant based on LLMs has also been introduced in Google Maps. LLM finds answers to user questions by analyzing more than 250 million locations on Google Maps and reviews from a community of over 300 million people. Naturally, it is expected to have a significant impact on the shopping experience at brick-and-mortar retailers. There are a growing number of examples of LLMs being actively leveraged to improve customer experiences both online and offline. The principles are the same. Among numerous shopping lists and location options, LLMs will help you find the answer that best suits your intentions and goals.
Allganize also helps companies immediately use the app that best suits their intention and purpose of AI adoption.
