Shopify goes all in on AI agents to revolutionize online shopping
Shopify is about to lead one of the most significant shifts in e-commerce in recent years. The company’s president, Harley Finkelstein, took the stage at the Upfront Summit 2025 in Los Angeles and left no room for doubt: the company is going all in on AI agents to redefine the way people shop online. And when you look at the context behind this decision, it becomes clear this is not just hype — the numbers and strategy back up the bet in a very concrete way.
Today, Shopify holds the position of second-largest e-commerce platform in the United States, trailing only Amazon. That is a standout position for sure, but what really grabs attention is the sheer size of the opportunity still ahead. According to Finkelstein, only 18% of retail purchases in the U.S. happen in the digital space. That means more than 80% of consumer spending still flows through brick-and-mortar stores, representing a massive territory waiting to be captured by e-commerce. And that is precisely where artificial intelligence steps in as the centerpiece of Shopify’s strategy for the coming years.
How AI agents will work in online shopping
Shopify’s proposal is to build AI agents that act as a kind of digital personal shopper — except far smarter and more accessible than any shopping assistant that exists today. These agents will be able to learn each consumer’s preferences over time, analyze thousands of products simultaneously, compare prices, evaluate reviews from other buyers, and recommend exactly what makes sense for each individual person.
The difference from traditional search engines is that the agent will not simply list generic results based on keywords. It will understand the context behind the consumer’s need and deliver truly personalized suggestions, taking into account purchase history, lifestyle, and even available budget.
Finkelstein used a practical example to illustrate this difference. He mentioned his personal preference for the sneaker brand On. When someone searches for sneakers on a traditional search engine, the results tend to surface large retailers like Footlocker at the top. The current system favors whoever has more domain authority, bigger ad budgets, and a stronger digital presence. But with an AI agent that already knows the consumer’s preferences, the logic shifts entirely. The agent knows that person likes On and will prioritize options from that brand, instead of directing them to mass retailers that dominate traditional results.
According to Finkelstein, the agentic model is fundamentally merit-based, unlike the current search model, which often favors those with more resources to invest in visibility. This distinction is central to understanding why Shopify believes so strongly in this transformation.
An important counterpoint about personalization
It is fair to point out that current search engines already do a reasonable job of personalization. Google, for example, adjusts shopping results and search rankings based on browsing history and previous user behavior. So saying that search engines do not personalize at all would be an overstatement. However, it is also true that the level of personalization AI agents promise goes far beyond what exists today. An agent that continuously follows the consumer, understands conversational context, and can cross-reference dozens of variables at the same time operates on a completely different level of sophistication. The question is not whether search engines personalize — it is how much further AI agents can push that personalization beyond what search engines can achieve with their current architecture.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
Beyond that, these agents will not just answer questions or display products. Shopify’s vision is for them to carry out complete actions, like adding items to the cart, applying discount codes, and even completing the purchase, all with the user’s approval, of course. This ability to act autonomously is what sets an AI agent apart from a simple chatbot. While a chatbot answers questions within a predetermined flow, the agent makes decisions, executes tasks, and learns from each interaction to continuously improve its recommendations.
The influence of social media and the authenticity of AI agents
Finkelstein also made an interesting observation about how AI agents differ from other forms of influence on consumer behavior. He acknowledged that people will continue to be influenced by figures they follow on social media and television, as has always been the case. The difference, he said, is that a chat-based app works as a more authentic personal shopper, precisely because it does not work on commission.
This is a point worth paying attention to. When a digital influencer recommends a product, there is almost always a commercial deal behind it — whether it is a paid partnership, an affiliate code, or a product exchange agreement. Consumers do not always have clarity on how genuine that recommendation actually is. An AI agent, in theory, will only show the consumer what it considers most relevant and most likely to meet their real needs. There is no commission involved in the recommendation, which could build a different level of trust in the relationship between consumer and technology.
Of course, this depends on how these platforms will monetize their agents. If at some point merchants can pay to have their products prioritized in agent recommendations, that promise of authenticity loses weight. But the original idea presented by Shopify is that curation should be based on real relevance, not on who pays the most to show up.
The impact for small and mid-sized merchants
Perhaps the most interesting aspect of this move by Shopify is not the technology itself, but who it could benefit. According to Finkelstein, AI agents have the potential to open a brand-new gateway into e-commerce, and this is especially relevant for small and mid-sized entrepreneurs.
Historically, these merchants have always faced a massive barrier to being discovered by consumers. Competing with large retailers in paid ads requires hefty budgets, and ranking in the top positions on search engines demands heavy investment in SEO and content marketing. With AI agents curating products based on relevance and quality — and not just on who pays the most per click — the discovery logic changes completely.
Finkelstein pointed out that many merchants on Shopify struggle to get their products discovered. And it is precisely in this scenario that he believes AI agents will play a critical role, surfacing new brands for the right consumers. When the algorithm works in favor of relevance rather than just marketing budget, small stores with quality products get a real shot at competing.
Think of it this way: if a consumer asks an AI agent to find a handmade leather bag crafted in Italy, with good value for money and positive reviews, the algorithm will not necessarily prioritize the most famous brand. It will search for the product that best matches those specific criteria, even if it comes from a small shop with only a few hundred sales. This democratization of consumer access could rebalance the playing field in e-commerce, allowing quality products to gain visibility regardless of the size of the operation behind them.
The virtuous cycle of AI-driven loyalty
Another point worth highlighting is the loyalty factor. When an AI agent learns that a particular consumer prefers sustainable brands, for example, it starts prioritizing merchants that fit that profile. This creates a virtuous cycle where smaller stores that invest in differentiators like sustainability, local production, or product customization naturally get recommended more frequently. In a scenario like this, competition stops being solely about price and logistics and starts valuing attributes that big retail chains often cannot replicate with the same authenticity.
What Shopify is building right now
The company is not just talking about the future — it already has projects underway. Finkelstein revealed that Shopify is developing an AI assistant for merchants called Sidekick. This tool will help sellers manage their operations more efficiently, using artificial intelligence to automate day-to-day tasks and provide business insights.
In addition to Sidekick, the company is also building a dedicated agent to handle customer support issues. The idea is for this agent to resolve service problems autonomously, reducing wait times for consumers and lightening the load on merchants’ support teams.
And there is more: Shopify is developing a protocol that allows AI agents to better understand merchant data, such as product details, inventory, variations, and technical specifications. This protocol works as a communication layer between the agents and store information, ensuring that the agents’ recommendations and actions are accurate and up to date. Without this well-structured data layer, even the smartest agent in the world would fail when suggesting a product that is out of stock or recommending a color variation that no longer exists.
What this means for the future of e-commerce
Shopify’s move is not happening in isolation. Companies like Google, Amazon, and numerous tech startups are also investing heavily in AI agents focused on consumer shopping. OpenAI, for example, has already announced its own agentic shopping system, competing directly with Google and Amazon on this new turf. What sets Shopify’s approach apart is the scale of its merchant network — millions of stores across multiple countries — and the fact that the company does not sell its own products. It functions as infrastructure for commerce, which puts it in a strategic position to serve as the connecting link between consumers and a vast variety of brands and products.
If the platform’s AI agents are well executed, Shopify could become the leading product discovery channel in the world, competing for space with Google search and Amazon’s marketplace. It is a massive ambition, but one that makes sense when you look at the current strategic chessboard.
It is worth remembering that this transformation will not happen overnight. Finkelstein himself acknowledged that initial adoption will be gradual. Implementing AI agents in online shopping depends on factors like consumer trust, payment integration, data privacy concerns, and of course, the actual quality of the recommendations. If the agents deliver generic suggestions or push sponsored products disguised as organic recommendations, the experience loses value fast.
But if Shopify manages to keep its focus on genuine relevance for the consumer, the growth potential is enormous — both for the company and for the entire network of merchants that relies on the platform to sell online.
A new era for digital commerce
Finkelstein wrapped up his appearance at the Upfront Summit with a level of enthusiasm you rarely see from executives at his level. He said Shopify is probably more excited about this new era of commerce than it has ever been in the entire history of the company. And the reason is simple: AI agents will not just create opportunities for large merchants. They will also benefit the so-called long tail of sellers — those thousands of smaller stores that currently struggle to be seen in the ocean of e-commerce options.
What becomes clear is that e-commerce is entering a new phase. The era of static catalogs and keyword searches is giving way to dynamic, conversational, and deeply personalized experiences. Shopify seems to have understood this before many of its competitors and is positioning itself to lead the transition. The concept of agentic shopping might sound abstract right now, but once the first real results start rolling in, the conversation about the future of digital retail will shift to an entirely different level.
For anyone selling online or thinking about getting started, keeping a close eye on this evolution is no longer optional — it is pretty much a necessity. 🚀
