Can AI and machine learning help Amazon make shopping easier and more natural?


“Machine learning is ubiquitous at Amazon today,” said Rajeev Rastogi, vice president of machine learning at Amazon India, in an interview with Gadgets 360. “Within retail, we use IT extensively. ‘machine learning to recommend products to customers, predict future product demand, and improve the quality of a product catalog, both by categorizing products and eliminating duplicate products.

One of the most basic examples of how Amazon uses machine learning (ML) is when you misspell a query in its search bar. The e-commerce site, Rastogi noted, examines the phonetic distance between the misspelled query and the correct query instead of looking at their text distance to provide accurate results whether or not you spelled something wrong.

For example, if you type “geezer” on Amazon to search for available geyser options, the marketplace will automatically correct the spellings and show you the relevant results. Amazon also uses ML models to translate content on its site into the Indian languages ​​it now supports.

Amazon uses machine learning to deliver accurate search results even if you misspell queries

Of course, these types of computer use are now commonplace, and it’s not something most of us think of when we consider the terms artificial intelligence (AI) or machine learning.

Rastogi revealed that his team is currently working on a startup initiative that aims to bring a conversational shopping experience. It is aimed at new online shoppers who are more used to communicating with offline merchants rather than placing an order through an e-commerce site.

Conversational commerce, through chatbots, through smart assistants like Amazon’s Alexa, is one such idea that comes up every few years as technology improves, and Rastogi explains how it’s going to start with text. , in English, but evolve into other languages, and by voice.

“A machine can read a document and then answer any question about the document, it’s difficult. Today, AI cannot generate reviews for a film, for example… Even summarizing a set of documents is a difficult problem. It is in no way solved by AI, ”Rastogi pointed out.

Rajeev rastogi amazon india computer scientist rajeev rastogi image

Rajeev Rastogi recognized the challenges of AI
Photo credit: Amazon

AI has been used to analyze text and speech at different levels. But computer engineers and data scientists have yet to find a relevant mix for using AI and machine learning to generate accurate ratings like movie or product reviews. In a research article, published by researchers Gerit Wagner, Roman Lukyanenko and Guy Paré of the Department of Information Technology, HEC Montreal, on how AI can be used in the literature review process, he is noted that even “technically perfect tools (like researchers)” sometimes have difficulty evaluating information from sources that use ambiguous and confusing language and presentation.

McKinsey Global Institute (MGI) partners Michael Chui, James Manyika and Mehdi Miremadi also pointed out in an article that AI models “have difficulty passing their experiences from one set of circumstances to another” and force them to do so. companies train models even when the use cases are very similar. This adds additional resource requirements.

Shreyas Sekar, assistant professor of operations management in the Department of Management at the University of Toronto Scarborough and the Rotman School of Management, said the effectiveness of an AI-based bot communicating with and giving humans appropriate results, especially in markets like India, is uncertain. . Sekar has done extensive research on how e-commerce platforms use machine learning on both the consumer and warehouse side to improve their operations.

“When you ask these chatbots simple questions, like no, is it going to rain tomorrow?” Or can you play me the song from that movie? They are doing a good job. But as you start to get more and more complex questions like “hey, can you help me find a good shoe for my trek?” I think it’s very difficult for the chatbot or even Alexa to clearly break this down into what is your intention? What do you do as a person and how do you differentiate yourself from others? And which products suit you? he said.

Manage bias and errors
One of the biggest challenges in using AI and ML these days is limiting bias and errors. Companies from Google and Facebook to Microsoft are regularly confronted with these blunders. Amazon is also not foolproof on this front.

Sekar of the University of Toronto Scarborough and the Rotman School of Management noted that Amazon’s AI deployments include many biases that the company is already aware of and apparently working to address, but doesn’t know in. how successful she was in achieving the desired results.

“For example, maybe historically users have clicked on a particular brand of headphones, and then what happens is that in the future I will continue to amplify that exact brand over and over again. . So it’s usually called a kind of popularity bias where I try to spotlight products that are already popular, and I basically help the rich get richer in the system, ”he said. .

Rastogi strongly disagreed, however, and said Amazon’s goal is to help human workers, not replace them entirely.

Who does it help?
Using AI and ML helps Amazon deliver what you need by understanding your buying behavior and buying history. However, this sometimes leads to impulse buying and just convinces you to buy something that you don’t actually need. Experts believe it would grow further with a more conversational shopping experience.

“I think AI and ML can definitely increase the idea of ​​converting storefront buyers into repeat buyers,” Sekar said. “And that’s definitely something that I think is a good way to view Amazon as a very persuasive seller.”

Amazon in the impulse buy image of the Amazon homepage

Amazon uses AI and ML algorithms to persuade customers to buy new products

Consumers themselves can overcome this behavior by understanding how algorithms can influence their choices.

“Even though we’re the one who clicks on a product to buy at the end, we’re sort of guided through the buying funnel by the algorithm in different places, whether it’s the recommendation or critics, ”Sekar said.

Ankur Bisen, senior partner and head of Consumer, Food and Retail divisions at management consulting firm Technopak, said the nature of how Amazon uses its algorithms to get consumers to buy more was exactly similar to this. as advertisements, marketing and even discounts at a retail store did.

“Amazon does it with a lot of precision because it’s defined,” he said. “Conversational AI is not just close to Amazon’s monopoly domain. Yes, they are very good at it thanks to Alexa. But you’ll see conversational AI emerge in different forms offered by other technology platforms.


Comments are closed.