Imagine using a chatbot that has been trained on the vast product catalog of the online retailer to assist you with your Amazon purchases. Rufus is a generative AI-powered shopping assistant designed just for Amazon users, so you can now do that. Currently accessible in the US, conversational search functions similarly to ChatGPT. Users can text the AI-enabled technology to receive personalized information while preserving context, allowing them to ask follow-up inquiries.
To make shopping more convenient and easy, Amazon has added a number of new generative AI-powered features to its store in recent months. Customers may immediately grasp consumer insights with the aid of its AI-generated review highlights, which present common themes from dozens, hundreds, or even thousands of reviews at a glance. With the help of Fit Review Highlights, a new feature from Amazon, users can now choose the perfect size by receiving individualized advice and insights. Additionally, it is utilizing generative AI to assist its selling partners in creating more captivating and effective product descriptions and titles, as well as to enhance already-existing listings, thereby increasing the amount of information available to consumers.
The widespread use of generative AI is not a mirage. According to Rajeev Rastogi, vice-president of machine learning at Amazon, “it is the next wave of deep machine learning (ML) adoption with the potential to reinvent customer experiences.” In order to enhance consumer experiences, the organization has been utilizing AI and ML extensively in recent years. Thanks to scalable technologies, Amazon has been able to optimize for India. “Amazon uses artificial intelligence (AI) and machine learning (ML) to validate customer addresses, compute address quality scores, label addresses as residential or commercial, fix city pin code mismatches, and provide suggestions to users to rectify wrong addresses, among other things,” stated Rastogi. She also added that this has helped delivery staff save hundreds of hours of address searching and fulfill delivery commitments on time.
According to Rastogi, “Amazon uses both proprietary and open source technologies to accomplish some of the generative AI applications.” To provide thorough product descriptions, the company uses large language models (LLMs), a kind of machine learning model that is specifically trained on massive volumes of data and is capable of text generation, summarization, translation, prediction, and recognition.
Rastogi claims that Amazon lists hundreds of millions of products in India alone, and that the company has billions of products worldwide. There may be problems with data quality when dealing with extensive product catalogs such as ours. He continued, “At Amazon, we have skilled data scientists and machine learning experts that are committed to solving problems with data quality.
