Shoppers already have agents. Does your store?

A purchase used to start on Google or on social media. Today, more and more people ask ChatGPT, Gemini or Meta AI first: what to buy, where it is cheapest, which one has the best reviews. And some already let an AI agent do part of the work for them.

That changes two things for a store: how its shoppers find it, and how it has to work so it does not fall behind.

An AI shopping agent is an assistant that searches, compares and, in some cases, buys for the person, asking them to confirm before paying. This model is called agentic commerce. In October 2026 it works mostly in the United States; in Latin America there are pilots with banks, and assistants already recommend products.

How shoppers use AI in 2026

In Latin America the change is already visible. In a Visa survey from October 2025, 75% of people in Mexico and 70% in Brazil said they had already used AI to help them shop, mostly to research products. The survey spoke with 1,000 people in each country.

In the United States there is traffic data. According to Similarweb, AI appears in 11.4% of the journeys that end at an online store, up from 4.5% in July 2024. Worldwide, visits that come straight from AI assistants grew 203% in a year, though they are still less than 1% of e-commerce traffic.

That is why we say “more and more shoppers”, not all of them. What is growing fast is AI’s weight in the decision, more than the traffic it sends.

What changes for your store

AI rarely replaces the rest of the journey. The shopper researches there, moves on to other sites and almost always uses a search engine too. Journeys that combine the two are the ones that most often end in a purchase.

For a store, that means three things:

• Show up where the shopper asks. If an assistant does not mention your brand when someone asks about your category, you are not in the conversation.
• Keep a catalog that makes sense. Clear names, descriptions, prices and availability, for people and for the AI that reads them.
• Do not drop Google. AI adds to search; it does not replace it.

Which AI agents already buy, and where?

The next step has already started: agents that do not just recommend, but buy for the person, always with their approval.

In September 2026, Meta launched Muse, a personal agent that works in its own app and inside WhatsApp. It can complete purchases, and before paying it asks the person to confirm. For now it is available in the United States. The same month, OpenAI introduced dots, agents that work continuously inside ChatGPT.

In Latin America, for now, these are pilots. In March 2026, Santander and Visa completed purchases made by AI agents in Argentina, Brazil, Chile, Mexico and Uruguay, in a controlled test.

For a store, the takeaway is practical: agents that buy are only starting to reach checkout, but the assistants that recommend are already deciding which products the shopper sees. That is where the work is today.

Why the store needs an AI operator too

While the shopper has an assistant that searches, compares and summarizes for them, most stores still run their operation across many tools: the store platform, Google Analytics, campaigns on Meta and Google, email, inventory and spreadsheets.

Pulling that information together takes time. In a 2025 BCG survey of 350 retailers, merchants spent 40% of their time on manual tasks that could be automated, and 69% of their leaders expect AI to free up time for strategy.

Budgets are moving too: in Gartner’s 2025 survey of 402 marketing leaders, 39% planned to cut what they spend on agencies.

The question is no longer whether to use AI in the store, but how: as one more tool someone has to check, or as someone who checks for you. To see how an operator differs from a chatbot or an assistant, read chatbot vs AI assistant vs AI agent.

What an operator like Tino does

An AI operator works on its own, on its own schedule, within the limits you set. It does not wait for you to ask.

Tino, the AI operator for online stores built by Merkanthub, connects to your store and to the tools you already use, such as Shopify, VTEX, WooCommerce, Google Analytics, Google Ads, Meta and Klaviyo, and even your ERP or your spreadsheets. With that, Tino:

• Reviews the business every day and tells you about the changes, risks and opportunities that matter.
• Walks through your store and tells you if something stopped working.
• Shows you how you appear in ChatGPT, Gemini, Claude and Perplexity, and what your pages are missing to show up.
• Builds collections from what is selling, at what price and with how much inventory.
• Prepares your reports on the schedule you set and writes to you on WhatsApp.

You decide what it can do on its own and what needs your approval. Just as the shopper trusts their assistant to search, you trust yours to look after the store.

Shoppers already have agents. Stores have Tino.

How to get started

1. Connect your store. It takes a couple of minutes, and from there you can talk to Tino.
2. Let it learn. Within hours it understands the basics, and within a couple of days it knows your store’s rhythm.
3. Set the limits. What it tells you about, what it prepares and what it can do on its own.

The fastest way to see it is on a real store: book a demo and we will show you what it found today.

Sources

Checked on October 6, 2026.

1. Visa, survey on AI and holiday spending in Latin America, fielded October 14–16, 2025 and published December 12, 2025.
2. Similarweb and Statista, State of Ecommerce 2026, September 2026. United States desktop data to June 2026, pp. 4, 9, 13, 29 and 67 of the report.
3. Meta, Muse announcement, September 2026.
4. TechCrunch, OpenAI launches dots, September 29, 2026.
5. Santander, AI-agent payments pilot with Visa, March 2026.
6. BCG, What 350 Retailers Say About the Future of Merchandising, May 12, 2025.
7. Chief Marketer, on Gartner’s 2025 CMO Spend Survey (402 marketing leaders), May 2025.