Emil Waszkowski
VP of Commercial Development
This time, the “e-commerce revolution” isn’t clickbait. Have you been searching for information in AI chats instead of Google lately? Your customers are doing the same – and it’s reshaping their shopping journeys. LLM providers already enable purchasing products directly inside a chat window, and AI agents are becoming shopping assistants that complete transactions on behalf of users.
With ChatGPT’s 800M weekly active users and the introduction of Instant Checkouts and Apps, it’s joining the ranks of the biggest consumer platform businesses. Google is rolling out AI Mode more broadly, and Amazon is openly showcasing how its AI features are driving sales.
This paradigm – called “Agentic Commerce” – forces e-commerce and retail brands to rethink how they reach customers, build loyalty, and optimize their sales processes.
Despite the huge volume of available shops online – or maybe precisely because of their overwhelming number – finding exactly what a consumer wants is more difficult than ever. Googling is often a fool’s errand, with the same ads conquering most keywords, and scrolling through numerous websites is time-consuming. And instead of leading to a successful purchase, consumers end up with ‘decision fatigue’ and don’t buy anything at all.
But now consumers have an alternative – they can get results in minutes instead of hours, all thanks to AI agents.
AI agents are software programs designed to accomplish tasks and goals on behalf of the users. They can reason, plan, and retain information, operating with enough autonomy to make decisions, learn from experience, and adapt over time. So for example, while chatbots can find specific clothing based on a photograph or an object based on a napkin sketch, an AI agent will go much further; being able to finalize a purchase, taking control of transactions themselves, and bypassing inefficient, slow checkouts.
Consumers won’t need an AI agent to decide which brand of chips goes best with a movie night. But they will want help choosing over-ear headphones that are great for classical music and have active noise-canceling.
That’s why the role of agentic commerce will differ depending on the type of product and its purchasing journey. The more options and parameters a customer needs to consider, the more valuable these assistants become.
64% say yes — as long as the AI agent includes at least one of the following features:
Agentic commerce isn’t a thing of the future; they’re emerging as we speak, with various companies racing to get the biggest part of the AI pie.
Google AI mode, using the combined powers of Gemini, can now do three things:
In practice, Google can now help consumers search for ideas, find matching products, and get notified when the price drops to a satisfactory level. And when searching for a “thin linen suit perfect for yachting”, consumers don’t have to count on the shops to use that exact keyword – they can instead trust the AI to understand the context. It can connect the concept of “linen” to “summer”, and “yachting” with “vacations”, meaning its search can be much wider, uncovering lesser-known brands and hidden gems.
With the increasing number of users using ChatGPT, the company has spotted a great opportunity: enabling sales instead of just answering “best fries in Katowice” type of questions. Now, users in the US can place orders in their favorite Etsy shops, and soon enough, they’ll be able to browse through SKIMS, Glossier, Spanx, and Vuori.
What does the purchase process look like in practice? Users, while asking for products (for example, “best coffee with a strong plum flavor” or “suitable jumper for hiking in freezing temperatures”), will get recommendations accompanied by a “Buy” button. If they like what they see, users will be able to finalize the purchase directly in the window, without the need to switch tabs. In that case, ChatGPT acts like the middleman in the process, passing information back-and-forth between the consumer and the merchant. As of right now, the only extra fee is applicable on the merchant’s side; users pay nothing.
Moreover, businesses get access to the Agentic Commerce Protocol, co-developed with Stripe, to build their own integrations, making the latest technology more accessible to all.
Perplexity offers US users a new feature called “Buy with Pro”, allowing to finalize a purchase right on Perplexity’s website or app, similarly to ChatGPT’s Instant Checkout. But instead of acting based on written queries, Perplexity can also provide recommendations based on photos. So if you see a person with a fancy backpack on the metro, you can snap a shot and ask Perplexity to look for the exact match, or at least for something similar enough.
Perplexity also has launched a program for merchants called Perplexity Merchant. And while users are promised unbiased recommendations, being part of this program does increase a brand’s overall visibility.
With over 9.7 million sellers worldwide, Amazon overwhelms consumers with its rich product catalog. If someone cares about finding the best deal that combines good quality with affordability, the time spent just browsing can skyrocket pretty quickly. To help make better decisions – which, in turn, may lower the return rates and increase profitability – the company introduced Amazon Rufus, an AI-powered shopping assistant.
Using the information from product listing details, customer reviews, and community Q&As, it can answer specific questions about how easy kitchen appliances are to clean or how a backpack performs in rain, removing the need to do any further research.
Amazon Rufus also comes with superb product recommendations that consider specific circumstances, like weather or use cases, similarly to Google AI mode. And to ease the shopping dilemmas even more, Rufus can quickly describe the differences between pizza oven types or running shoes.
In case a consumer is unsure of what they need, they can simply describe the situation they found themselves in and count on Rufus to make appropriate recommendations, from “What to get on a Father's Day” to “What do I need for a garden party.”
Due to its success, Amazon estimates that Rufus will generate an additional $10 billion in annual sales for the company.
Amazon’s other feature also made the list – Help Me Decide. Users can count on the button of the same name to make browsing products easier than ever. For example, if a user selects the button while comparing different hoovers, Amazon’s AI will consider this specific’s user past behavior. If they were looking for pet accessories the day before, AI will connect the dots and recommend a hoover designed with pet hair in mind, along with a turbo brush and a HEPA filter to boot – something the user might not thought of before.
If you’re interested in implementing a similar feature in your e-commerce business, reach out to Future Mind’s representative to get a quote or see our case studies for brands like Sinsay, BIG STAR, Modivo.
Zalando, one of Europe's largest online fashion platforms that pioneered free shipping and returns policy back in 2008, partnered with OpenAI to create an AI-driven tool designed to deliver personalized content recommendations and simplify product discovery. The challenge lay in ensuring Zalando Assistant could hold nuanced conversations in 25 languages, but ultimately, it was worth it – resulting in a 23% increase in product clicks and over 40% more products added to wishlists.
Similarly to other examples, consumers can describe their situation – “What to wear for a wedding garden party in the middle of summer in Spain” – and then guide the Assistant to narrow down the search (“Short sleeves only, please”).
To ensure the chosen product fits like a glove, consumers can use the Virtual Fitting Room, where their 3D Avatars of the exact same measurements can try on anything they’d like. This immersive experience doesn’t only boost engagement – it also entertains.
Walmart Sparky, similarly to other solutions, Sparky speeds up the decision-making process by leaps and bounds. Users can come up with complex scenarios for Sparky to consider – such as “I need a rice cooker suitable for a family of five” – and the AI will browse through product descriptions and reviews to find the best possible choice; for example, a model with at least 1.8 liters of capacity, a keep-warm function, and a removable nonstick bowl.
Mastercard Agent Pay, built on tokenization capabilities in collaboration with Microsoft Azure OpenAI Service and Microsoft Copilot Studio, allows to safely and securely finalize any payments directly in AI chat conversations.
This means that a consumer can now ask the AI agent to find the best Xbox game for his nephew, adjusting the recommendations based on favorite genres, past purchases, current deals, and trending titles. In the same conversation, the AI agent can suggest the best way to pay and quickly make the purchase with no fuss or checkout struggles.
To help avoid fraud and ensure security, the AI agents use on-device biometrics to confirm purchases and clarify agentic transactions that may be unfamiliar or unrecognized.
Sierra is a real catch for companies that want to step up their customer service game. While it doesn’t work as a substitute for human employees, it can help out in specific, repetitive scenarios – such as cancelling tickets, finding out the state of an order, or updating a subscription.
This solution isn’t just an advanced chatbot; it’s a platform for designing whole AI agents, specialized in their niche, that can moreover keep learning and improving their craft. The AI can significantly reduce the number of tickets human employees have to deal with, leaving them the more complex cases, while ensuring that each and every client gets prompt replies – and in brand’s tone of voice, too.
To ensure that AI won’t provide false information, Sierra uses supervisory layers to reduce hallucinations and prevent abuse.
Firmly is an agentic commerce platform that doesn’t allow brands to waste any opportunities. Ads, social media posts, blog posts, chatbots, SMSs, streaming videos, and marketplaces – they all can be used to sell products thanks to instant checkouts. And that’s exactly what Firmly offers; a seamless integration with leading digital wallets throughout various channels.
Hard to imagine? Consider this: a user comes across an interesting book promoted in an Instagram post. If it catches their attention, the “Place Order” button can make them consider making a purchase without ever leaving the platform. A whole journey from Googling a product, comparing prices, and going through checkout can be now shortened to just a few clicks in one tab.
The Universal Commerce Protocol UCP is an open-source standard that gives AI agents and commerce systems a shared language for the entire shopping journey. From discovering products and checking inventory to managing carts, completing checkout, and handling post-purchase updates like order tracking, UCP makes AI interactions seamless and consistent.
Google built UCP together with Shopify, Etsy, Wayfair, Target, Walmart, and more than 20 other partners. The goal is to let AI agents talk directly to merchant systems without slowing anyone down. Brands stay in control of pricing, business rules, and customer relationships, while AI can take care of the repetitive steps in the shopping journey.
This also opens the door to AI-assisted shopping across platforms like Google Search AI Mode and the Gemini app. Developers and payment providers can use UCP to create experiences that work seamlessly across multiple agents and stores.
UCP makes agentic commerce work by streamlining interactions, ensuring security, and giving AI the tools to assist shoppers from beginning to end.
While the future is difficult to predict, we already see the direction household names are taking as they further develop their AI agents and embrace agentic commerce. We can expect hyper-personalized shopping, where AI analyzes browsing behavior and past purchases to deliver real-time recommendations at scale. We’ll also see more frictionless store experiences, with AI-powered checkout systems that remove wait times, reduce fraud, and enable seamless, secure transactions.
It’s quite possible that shopping may shift from an active decision-making process to a largely passive, behind-the-scenes automation. The traditional experience of browsing, searching, and selecting products could be replaced by AI proactively making choices for consumers, streamlining retail while fundamentally changing the nature of consumer engagement.
From a business perspective, agentic commerce introduces both opportunity and risk. It’s a chance to gain visibility, but it might also turn into a struggle against an algorithm. One thing’s clear: businesses that adapt quickly and use AI thoughtfully will be best positioned in a retail landscape shaped by agentic commerce.
A. Rakowski - Partner Relationship Manager
P. Wasilewski - Head of Solutions
E. Waszkowski - VP of Commercial Development
P. Wasilewski - Head of Solutions
