Amazon recently announced its plans to launch a generative AI tool destined to enhance search within its ecommerce platform. This announcement didn’t come through a press release… but via job postings, which are often the best ways to gauge a company’s ambition. This comes as the battle for AI search heats up between tech giants.
One such listing describes the ideal candidate as someone being able to “reimagine Amazon Search with an interactive conversational experience”, while another speaks of “a new AI-first initiative to re-architect and reinvent the way we do search through the use of extremely large scale next-generation deep learning techniques.” “This will be a once in a generation transformation for Search.” No pressure.
We’ve seen how powerful large language models can be; ChatGPT-like tools could transform Amazon’s e-commerce platform, probably for the better. Its search bar has become the go-to gateway for millions of shoppers, with more than half of US shoppers beginning their product searches on Amazon.com, surpassing Google. Meanwhile, the company has been heavily criticized in recent years for the increased share of results devoted to ads and other sponsored content.
Three Opportunities for Generative AI at Amazon…
1. A Conversational Shopping Experience
Browsing through Amazon’s extensive inventory can often be overwhelming, especially when looking for a specific item or exploring new categories. Using a generative AI solutions “in” the e-commerce search bar would enable a more natural shopping experience. One could write “Show me some cozy sweaters for winter in New York” or “I took French for 2 years in high school, what are some good books for learning French?”, and get results taking into account the average winter temperatures in New York, or French books that do away with very basic introductions.
There is a world in which Amazon develops an experience akin to engaging in a conversation with a knowledgeable salesperson (who would just happen to be called Alexa, I guess). This conversational approach would empower users to navigate through the vast sea of products with ease, saving time and reducing frustration.
This will be a significant departure from traditional keyword-based searches. We’ve gotten so used to typing very specific words in a very specific way, that conversing more naturally while online shopping will take some getting used to.
2. Personalized Recommendations
Since Amazon has a full history of shoppers’ previous purchases, results would likely also take into account personal preferences (or at least use them as part of the conversation). By analyzing past shopping behavior, search history, and user feedback, Amazon’s AI-powered system would make tailored recommendations that suit a customer’s unique tastes and requirements.
Some will argue that this level of personalization not only enhances the shopping experience, but also enables users to discover new products and brands that align with their interests, fostering a sense of discovery and delight. I argue that Amazon has used AI to make recommendations for years, and they’ve always been awful. But hey, maybe more tech is the answer.
3. Conversational Customer Support
With ChatGPT-style search, Amazon probably envisions extending its conversational capabilities beyond the realm of product discovery. The system has the potential to serve as an intelligent customer support assistant, addressing queries, providing product information, and offering guidance in real-time.
Customers could seek assistance using natural language, eliminating the need to navigate through extensive FAQ sections or wait for a response from traditional customer service channels. This seamless integration of AI technology may enhance customer satisfaction and streamline the support process. Which is great, as long as it does not cost jobs, but rather improves them.
… and Three Challenges
Of course, there are also some challenges and risks involved in using generative AI for online shopping.
1. The Privacy & Safety Concerns
For one thing, the technology is far from perfect, sometimes producing inaccurate, misleading, or even harmful content. For example, it might suggest Amazon products that are out of stock, unsafe, or even illegal. Or it might generate biased or offensive text that could offend customers or violate regulations.
Furthermore, a ChatGPT-style search would rely on user data to provide personalized recommendations. Amazon must prioritize privacy and ensure that customer information is protected and handled responsibly. Transparency and user consent should be at the forefront of engineers’ minds to maintain trust and safeguard sensitive information.
2. Conflicting objectives
Another challenge is how to balance the interests of customers, sellers, and Amazon itself. By leveraging sophisticated algorithms and personalized recommendations, generative AI has the ability to manipulate customer behavior, nudging them towards buying more products or spending more money than originally intended.
Amazon will also need to avoid any bias or favoritism towards certain sellers or brands, as this could undermine the trust and fairness that customers expect from the platform.
To address these concerns, the company must ensure that its generative AI system operates with the utmost transparency, allowing customers to understand how recommendations are generated and giving them control over their own shopping experiences. Implementing stringent ethical guidelines and mechanisms for auditing and monitoring the AI algorithms will be crucial in maintaining fairness and mitigating potential risks.
3. Doing away with humanity
Amazon implementing AI to improve search is the result of a long-standing (if unofficial) philosophy within the company : More Tech Is The Answer. It’s in their DNA. But one may wonder if there could have been a more humane way to solve the issues the company is facing with search.
Exploring alternative avenues, Amazon could have delved deeper into the realm of retail by creating physical spaces where customers could interact with knowledgeable salespeople. By offering a personal touch, these encounters would have bridged the gap between virtual and real-world experiences, facilitating a deeper level of engagement and understanding between consumers and the products they seek.
Furthermore, the company could have considered augmenting their workforce by hiring additional personnel dedicated to enhancing the search process. While this approach may incur higher costs initially, it could have fostered a more human-centric approach to search, leveraging the expertise and intuition of real individuals to complement the capabilities of AI technology. After all, the human touch can often provide valuable insights and nuances that algorithms alone may struggle to capture.
The principals of capitalism will argue that more growth requires more tech, because people are expensive. But I don’t know how long this trend can continue before society collapses on itself.
Amazon’s venture into ChatGPT-style search signals a potentially exciting future for online shopping. By harnessing the power of conversational AI, the e-commerce giant will surely create a more personalized, intuitive, and engaging experience for its customers.
But, as we embrace this new era of intelligent product discovery and conversational interactions, it is crucial to remember what these new technologies may end up costing us.
Good luck out there.