Contact us!

How to sell on Shopify when the purchase starts in AI

By /

The conclusions of the the webinar we delivered on June 4, 2026 at the Madrid Chamber of Commerce, together with Monei and as part of the TIC Negocios program, distilled into 5 pieces.

Imagine a customer who wants to buy matcha tea. They no longer open Google or type in your website’s search bar. They open ChatGPT and writ “recommend me a premium matcha tea to drink every day” AI replies with three stores, photos, prices, and a link to buy. The uncomfortable question is simple:

When your customer asks an AI about what you sell, do you show up… or does your competition?

This is the conversation we are having most often at Impulsa3 with brands that have been selling well through their ecommerce for years and, suddenly, notice that discovery traffic is moving to a new place where they don’t know whether they are present. And it is almost never a product problem. It is that the starting point of the purchase has changed, and the store has not adapted to that change.

That was exactly what we put on the table in the webinar, with four complementary perspectives: visibility (Jorge Nevado, CEO of Impulsa3, store and development (Javier Hernández), integration and data (Miguel Martínez, CTO) and payments (Alex Saiz, CEO of Monei). Here is the full recording:

What has changed is not a tool: it is where the purchase begins

For twenty years, selling online meant appearing on Google and converting traffic on your website. That model is still alive, but it is no longer the only entry point. Today, a growing share of decisions starts in a conversation with AI —ChatGPT, Microsoft Copilot, or Google Gemini— that discovers, compares, and recommends products before the user sets foot in a single store.

The data we shared in the session shows the scale of the change:

  • 35% of consumers already use AI to discover products, compared to 13.6% who start from a search engine (Similarweb, 2026).
  • Traffic arriving from AI to Shopify stores has multiplied by 7Traffic arriving from AI to Shopify stores has multiplied by 11, since January 2025 (Shopify, 2026).
  • Between the 12% and 18% of informational queries already go through AI in 2026; a year ago, it was less than 2%.

The operational conclusion is not “build yourself a chatbot.” It is that the ecommerce that wins in 2026 works like a system of five interlocking piecesvisibility, conversion, integration, payment, and the whole system working together. Let’s go piece by piece, with the specific actions that can be applied.

1. Making sure AI finds you: visibility in the age of GEO

The good news is that the customer is still yours. AI acts as a discovery channel: it brings the buyer to your product page and the sale is completed in your store (the parameter utm_source=chatgpt that you will already see in your analytics confirms this). The question is no longer “Do I rank first on Google?” but “How many times does AI cite me and recommend me?”

This is where the concept of GEO (Generative Engine Optimization): if traditional SEO fought for the number-one position, in AI visibility is frequency —appearing in many different answers—. And there is one detail that surprises almost everyone: around 87% of the sources cited by ChatGPT match Bing’s organic top results (Seer, 2026). The search engine many had written off as dead turns out to be one of the gateways to getting AI to recommend you.

The plumbing, without technical jargon: MCP, UCP, and AP2

In the session, we explained the three acronyms you are going to hear all year, in plain language:

  • MCP (Model Context Protocol, by Anthropic): the “universal plug” that connects AI with external tools and data. Every Shopify store already comes with it built in.
  • UCP (Universal Commerce Protocol, by Google and Shopify, January 2026): the “common language of commerce,” so that any AI and any store can understand each other without programming a custom integration for each platform.
  • AP2 (Agent Payments Protocol): the “payment notary,” which signs and authorizes a purchase when the user authorizes AI to buy on their behalf.

This ties into the most repeated question from the webinar, which deserves a direct answer:

Do I have to integrate via API with each AI to appear as a recommendation? No. So that you be cited, AIs read their own index (Bing in ChatGPT and Copilot; Google in Gemini): it is enough to be well positioned and have clean data. So that they can buy, Shopify already publishes for you the files that AIs use to discover you, and activating your catalog is a single switch in the admin (Agentic Storefronts), not a connection for each platform. Avoiding those one-by-one integrations is precisely what the new standard solves.

Now, the honest nuance: activating the catalog makes you eligible, it does not make you guarantee the recommendation. Each AI decides who gets recommended based on its own criteria. That is why what really moves the needle is still the same as always, done well: complete product pages, real reviews, loading speed, and descriptive titles (not just brand names). The schema schema helps —it improves visibility by up to 67%—, but on its own it is not enough.

How to know if it is working

  • Look at which of your URLs the LLM cites and run them through the structured data test.
  • Bing Webmaster Tools now has an “AI Performance” report (beta) showing the pages from your domain that AI cites.
  • Google Search Console launched its generative AI report in June 2026, with impressions in AI Overviews and AI Mode.

2. Making sure your store converts and sells in every market

There is no point in AI bringing you the customer if your store loses them. And a lot is lost: more than 70% of carts are abandoned at checkout, and around 20% of those abandonments are due to the user experience (Baymard).

Shopify has embedded AI inside the dashboard itself —Sidekick performs tasks, writes campaigns, and creates segments by speaking to it in natural language— and has added tools such as A/B tests for themes native. It helps a lot, but it is worth knowing its limits: AI can generate themes, sections, and apps, but it does not guarantee consistency, maintainability, or the integrity of your data. The quick “do it yourself” and the sustainable “do it properly” are not always the same, especially when apps live on third-party servers and Shopify has moved from two major releases a year to major changes almost every month.

A new feature that is especially interesting for SMEs: selling B2B (company profiles, tailored pricing, volume discounts, 30-day payment terms) and operating in multiple markets (catalog, prices, language, and look and feel by country, from a single store) is no longer exclusive to the Plus plan. Since 2026, it is also available on Basic, Grow, and Advanced, with their respective limits. In other words: selling to companies and across multiple countries no longer requires Shopify’s most expensive plan.

3. Making sure everything is connected, migrated, and measured

Your store is not an island: the same sale exists at the same time in stock, in the ERP (the system that manages inventory, orders, and invoicing) and in the CRM (your customers’ system). When those pieces do not talk to each other, overselling, outdated stock, and returns turned into an accounting mess start to appear.

In 2026, Shopify expanded native ERP connectors, B2B store credit by company, and VAT ID validation (VIES) with reverse charge in checkout. Even so, real-world cases —packs, bundles, multi-warehouse, summary invoicing, channel-specific pricing— often require a serious connector with support. This is exactly the field of our Shopify Connector, which connects Shopify with your ERP/CRM without you having to touch anything manually.

And what if you also need to migrate from another platform? The three usual fears can be neutralized with method: not losing SEO (well-planned architecture and 301 redirects), not losing history (orderly transfer of orders, customers, and non-standardized fields) and not dropping during the transition (a planned and tested switchover window). A well-executed migration is not noticeable in sales: that is the test.

4. Making sure payment closes the sale (where it is won or lost)

You have managed to get AI to find you, your store to convert, and everything to be connected. And then comes the moment of truth: payment. In Spain, 3 out of 4 carts are abandoned before paying. And it is almost never because of the product or the price: it is because of the payment —distrust when entering card details, too many steps, or the customer’s favorite payment method is missing—.

In this section, Alex Saiz (CEO of MoneiIn this section, CEO of Monei) explained why payment also changes with AI. When the customer buys inside an assistant, four actors are involved with very clear roles: the AI platform (discovery and conversation), your business (catalog and, most importantly, merchant of record), the credential provider (Google Pay, Apple Pay…), and the PSP or payment gateway —Monei’s role—, which processes and settles the charge. The reassuring idea for any business: UCP does not turn you into a marketplace. PCI compliance, refunds, chargebacks, and strong customer authentication (SCA / PSD2) remain your responsibility, exactly as they are today.

And one very Spanish detail: there is more than 12 million Bizum users, which collects in less than five seconds and with no chargebacks. Shopify Payments does not include Bizum natively in Spain; a specific payment gateway is required, and Monei was the first one compatible with Bizum for Shopify. In a country where the payment method decides purchases, that is direct conversion.

5. The complete system, working together

If you take away just one idea, let it be this: the ecommerce that wins in 2026 is not the one with the best piece, but the one with all five pieces fitted together. AI finds you, your store converts and sells in every market, everything is connected and measured, and payment closes the sale. A failure in any link is reflected in revenue.

The usual trap is solving one piece in isolation —activating the AI catalog but with poor product pages, or having a beautiful store with mismatched stock, or a flawless checkout that AI never sends anyone to—. The value lies in looking at the whole system.

Where to start

At Impulsa3, we work on the four pieces under one roof —technology, marketing, and AI— as Shopify Partners, with consulting before execution and our own product for the integration side. If after the webinar you are wondering where your store stands, we can help you figure it out quickly.

We offer you a free diagnosis of your ecommerce: we review your visibility in AI, your store’s performance and usability, and give you back the 2–3 priorities that would have the greatest impact on your sales. With no commitment and with a business-driven approach, not just a technical one.

Request your free ecommerce diagnosis

Watch the full webinar.

How to sell in ecommerce when the purchase starts in AI (Shopify 2026) · Madrid Chamber of Commerce · June 4, 2026 · YouTube video

Speakers: