Generative AI is rapidly transforming the way consumers discover, evaluate and purchase products online. From personalised recommendations and AI-powered shopping assistants to multilingual conversations and predictive deal tracking, the technology is reshaping every stage of the customer journey.
In this interview with BW Businessworld, Kishore Thota, Director – Shopping, India and Emerging Markets, Amazon, discusses how Amazon is embedding AI across its shopping ecosystem, why hyper-personalisation is becoming the new baseline, and what the future of ecommerce could look like.
AI is increasingly influencing online shopping decisions. How is Amazon seeing customer behaviour evolve with tools such as Rufus, AI-generated product summaries and personalised recommendations?
We look at generative AI shopping through three key layers. The first is what we call the memory layer. It continuously learns from a customer's shopping behaviour, preferences and interests to create a unique shopping profile. The second layer consists of large language models (LLMs) that power intelligent shopping experiences. The third is the customer interface, whether that's conversational shopping, AI-generated insights or contextual recommendations.
Initially, we introduced AI-generated review summaries that condensed thousands of customer reviews into a few useful highlights. Today, those capabilities have evolved into what we call Quick View—a personalised product summary that doesn't simply describe a product but explains why it may be relevant to an individual customer.
For example, if you're shopping for earbuds and we've learnt that you're a regular runner, the summary may explain how securely the earbuds fit during a run. If you're also a swimmer, it may tell you they're not waterproof and therefore may not be suitable for your needs. That is where hyper-personalisation begins to create real value.
Beyond that, AI also interprets price trends. Our Price History feature shows how prices have moved over 30 or 90 days, while AI helps customers understand whether it is actually the right time to buy.
We also have Lens AI, which allows customers to simply point their phone at a product to identify and purchase it. In a market like India, where describing products is not always easy, visual search has become one of our most frequently used shopping features.
Looking ahead, I see two major shifts. First, customers will increasingly expect hyper-personalisation as a default. They will assume Amazon already understands what they need.
Second, shopping will become far more conversational. Instead of typing "red running shoes size 9", customers will naturally ask, "Help me find shoes for my morning run." AI will understand where they live, the terrain they typically run on and recommend the most suitable option while explaining why. That is the direction shopping is heading.
One of the biggest challenges in ecommerce is information overload. There are too many reviews, specifications and choices. How is AI helping customers make quicker and more confident decisions?
That is exactly one of the biggest problems we are trying to solve.
Over the years we have continuously added more information to help customers—videos, 360-degree product views, customer images, specifications, Q&A sections and reviews. But today some products have more than 20,000 reviews. That is almost like reading a 300-page book before making a purchase.
AI changes that completely. Instead of expecting customers to process enormous amounts of information themselves, AI reads everything, understands customer preferences and distils all of that into a few highly relevant insights.
Rather than producing a generic product summary, AI creates a personalised recommendation explaining why that product suits a particular customer. We are already seeing the impact. AI-generated product summaries have delivered a 33 per cent higher click-through rate than previous information formats in the same location.
Our ambition is not to replace existing information but to provide customers with the single most trusted starting point for decision-making. Of course, this comes with enormous responsibility. The recommendations have to be accurate because trust is earned only when AI consistently gets it right.
Trust has always been central to online shopping. How are AI-generated product insights and features such as Price History strengthening consumer confidence, particularly during events like Prime Day?
Trust operates across multiple layers. The foundation is accurate catalogue information. We use generative AI to help sellers create richer and more accurate product listings from the beginning.
The second layer comes from our customer review ecosystem. We have invested heavily in moderation systems that ensure reviews remain authentic and trustworthy. AI helps summarise thousands of reviews into concise insights without losing their essence.
Then comes pricing transparency. Our Price History feature shows price movements over both 30-day and 90-day periods, allowing customers to judge whether the current price genuinely represents a good deal.
We also display product velocity, showing how frequently an item is being purchased. While that does not necessarily determine buying decisions, it provides valuable social proof.
During Prime Day, these features become even more meaningful because customers can clearly see whether a Prime deal represents the lowest available price rather than relying solely on promotional messaging. Ultimately, transparency builds trust.
India is an extremely diverse market with multiple languages and varying levels of digital literacy. How is Amazon using AI to make shopping more inclusive?
There are two dimensions to that. First, AI has dramatically improved our multilingual capabilities. Content translation, catalogue localisation and campaign creation are now much faster, more accurate and significantly more cost-effective.
Second, generative AI is inherently multilingual. A customer can begin a conversation in English, switch to Hindi and then continue in Telugu without losing context. The AI understands intent across languages and keeps the conversation flowing naturally.
For India, where language diversity is immense, this capability is fundamental rather than optional. We are pleased with the progress we have made, but there is still much more to come.
Looking ahead, what role will generative AI play in reshaping ecommerce over the next few years?
Generative AI will be at the centre of shopping innovation. Our focus will continue to be making shopping easier, faster and more rewarding.
You will see conversational shopping becoming richer and increasingly proactive. Instead of customers always initiating interactions, Amazon will begin anticipating their needs. For example, before Prime Day we could proactively highlight products a customer has been tracking, notify them about the best offers, recommend the most suitable payment method and calculate the maximum savings automatically.
Customers should not have to work out which combination of coupons, bank offers or discounts gives them the best value. AI should do all of that. Ultimately, our goal is simple: reduce effort, improve value and make every shopping journey more intuitive. That journey has only just begun. |
