Where are we in the AI hype cycle? The initial shock of ChatGPT has subsided, and AI cynics now argue that progress in large language models (LLMs) has plateaued. Some argue that the hype is morphing into disillusion. In my opinion, this perception misses what’s most important.
While big tech may be struggling to show immediate profits from AI investments, it’s not because the innovations are ineffective. Rather, the underlying technology innovation is outpacing the rest of the business’s ability to add value while achieving economies of scale – reminiscent of the early days of the dotcom cycle, when the likes of WebVan went bust, only to materialize a decade later as successful businesses like Instacart and Amazon Fresh.
Soon, businesses’ ability to deliver value with these innovations will catch up. But we cannot wait for that shift. AI is growing more powerful by the day, shifting consumers’ fundamental expectations of technology. This is particularly true of the e-commerce sector.
When we consider the early days of the internet, Amazon gets the most credit for establishing the model of online e-commerce we have all now come to expect. The way we engaged with a search bar, the way we filtered and faceted products – we learned to shop this way because Amazon (and eBay, and even Google on a broader level) taught us that it was the norm.
That same shift is now happening again, this time driven by generative AI. Consumer interactions with ChatGPT, Gemini, and the like are instilling an expectation of conversational modes of search. Natural language is becoming more prominent in searches that were once limited to stilted keywords. That expectation is quickly going to translate to e-commerce sites, wholly transforming what it looks like to search and shop for your favorite brands.
For businesses to ignore this shift – to give in to disillusionment – is a disservice to their customers. AI is going to change e-commerce search. That’s not “hype,” it’s an inevitability … one that’s already starting to take shape.
The Power of AI-Driven Ranking
The shift to understanding natural language search queries is an area where this shift is already becoming evident. When it comes to longer searches, e-commerce search engines have long struggled to deliver the perfect results. Even the best legacy search technology has difficulty balancing recall (quantity of results) with precision (relevance). Increasing one typically means sacrificing the other. This limitation stems from the complexity of understanding shopper intent and the human language and can leave customers frustrated and unable to find the right products. For example, a search for “blue running shoes for bad ankles” might return irrelevant results like blue ankle socks, or too few results to choose from, potentially preventing conversion.
It is only with new innovations in AI that e-commerce search is gaining the ability to balance recall and precision. With a hybrid approach combining both semantic keyword technology and newer LLMs able to generate vector embeddings into a single search engine, customers can shop with natural language queries and still get highly relevant results. Search quality becomes a highly automated process – no matter the length or complexity of a query, AI can provide the right results. While we’re not yet at the point of perfection, AI is certainly redefining what’s possible in the search bar.
Advancing AI Search in E-Commerce
Modern AI-powered approaches even allow e-commerce teams to take this a step further, fine-tuning their search algorithms based on their unique data sets and on-site performance metrics. Instead of relying on predetermined rules, these advanced systems use machine learning (ML) to understand the complex relationships between products, user behavior, and purchase patterns. Now, the same “blue running shoes for bad ankles” search can prioritize results based on the shopper’s browsing history, previous purchases, and current trends, dramatically improving the likelihood of a successful transaction and a connected shopping journey.
Moreover, these AI systems can rapidly adapt to shifting market trends by continuously analyzing aggregate search and purchase data. This allows the technology to quickly identify emerging product preferences or seasonal shifts, which helps ensure that search results remain relevant even as consumer tastes evolve. This customization enables businesses to create highly personalized shopping experiences, significantly boost conversion rates, and drive revenue growth.
As AI advances search quality with more automatic optimizations, seamlessly adapting to shifts in customer search expectations, it will also leave more room for merchandising teams to explore business-specific fine-tuning. They can introduce custom signals for products, which are particularly valuable for new items lacking historical data. They can also adjust the weighting of various factors in ranking algorithms, balancing elements like product popularity and query relevance. For larger operations, there’s even the possibility of integrating proprietary ranking models.
Importantly, AI-powered search tools enable businesses to scale their operations efficiently. By automating complex decision-making processes and allowing for quick adjustments across large product catalogs, retailers can manage and optimize search results for millions of items with the same effort previously required for just a few thousand. This scalability allows businesses to expand their product offerings or enter new markets without proportionally increasing their workforce.
Neither Hype Nor Disillusionment
We have moved past inflated AI expectations into a phase of practical implementation and ongoing experimentation, where AI-powered search is actively reshaping online retail. Despite ongoing challenges, the ability to fine-tune search algorithms with ML and adapt to market trends in real time is driving unprecedented results.
Businesses are constantly in a state of experimentation, testing new approaches to balancing user preferences with business objectives. While still evolving, these AI-powered tools empower retailers to deliver hyper-personalized and intent-relevant shopping experiences. This positions AI in e-commerce not at the peak of hype but at the point of delivering value and sustaining innovation.