How AI is Revolutionizing Retail: A Conversation with Rama Ramakrishnan
In 2010, Rama Ramakrishnan had a realization while speaking with retail executives. He noticed that while personalized recommendation systems were gaining attention, they didn’t always benefit retailers. Additionally, most customers only shopped once or twice a year, so companies didn’t have much information about them. Ramakrishnan, a professor at the MIT Sloan School of Management, realized that by carefully noting customer interactions, a detailed picture of their preferences and behaviors could be created. This information could then be used with machine learning algorithms to improve the shopping experience.
Inspired by this idea, Ramakrishnan founded CQuotient, a startup that developed software for Salesforce’s AI e-commerce platform. Today, CQuotient’s technology interacts with billions of shoppers on Black Friday alone. After a successful entrepreneurial career, Ramakrishnan returned to MIT Sloan in 2019 to teach students how to practically apply AI technologies in the real world.
Ramakrishnan also participates in MIT executive education, where he learns about the concerns of senior executives. One common concern is the need for large amounts of data to train AI systems. Ramakrishnan is able to guide them to pre-trained models that can be quickly adapted to their specific needs.
AI, or artificial intelligence, is the quest to give computers the ability to perform tasks that only humans could do in the past. The traditional approach to AI involved applying if/then rules learned from humans. However, this approach had limitations. Humans can perform tasks effortlessly without being able to explain how they do them. This made it challenging to teach computers to do the same tasks. Machine learning, on the other hand, takes a different approach by allowing computers to learn from examples. This approach has been successful in tasks such as credit scoring, loan decision-making, disease prediction, and demand forecasting.
Machine learning worked well with structured data, but struggled with unstructured data such as images and audio. However, around 2010, deep learning emerged as a solution to this problem. Deep learning, based on neural networks, became practical due to the availability of large amounts of data, powerful processing hardware, and advances in algorithms and math. Deep learning has led to the development of generative AI software that can create realistic outputs, such as human-like text and images.
Generative AI, trained on vast amounts of text data, has the ability to predict the next most likely word and generate coherent text. However, it’s important to remember that the output may not always be accurate or relevant. Users have the responsibility to verify and correct the output before using it.
When applying generative AI in corporate settings, it’s important to consider costs and potential consequences. If correcting the generated content is less expensive than creating it from scratch, it might be worth exploring. Companies should also have a human in the loop to ensure the quality and accuracy of the content.
Currently, the most mature application of generative AI is in software development, where the technology can write code quickly. Software development already has infrastructure for testing and verifying code, making it a perfect fit for generative AI. Other applications include content generation and document search, but companies must be cautious when using generative AI in customer-facing chatbots.
Overall, AI is revolutionizing the retail industry by enabling personalized recommendations and improving the shopping experience. With advancements in machine learning and generative AI, retailers can better understand their customers and provide tailored solutions.