Eating your margin for breakfast: AI-shopping

The new shopping integrations between platforms like OpenAI and Shopify are super–exciting. They reduce friction like we have never seen before. Customers can discover products, ask questions, and complete purchases without leaving the conversation. It’s cool, from the consumer perspective. I have started using it on a daily basis. It’s good.

For many brands, it will feel like a new acquisition channel. Sounds great.

But there is an important strategic question underneath the excitement. It’s not about traffic. But about ownership.

Many product managers, brand managers and business owners spend a lot of time optimizing small percentages in their cost structure. They negotiate 0.5% savings with their supplier for scale, or just to see where the power is. They move payment processors to a competitor to save a fraction of a percent. Digitally, they run endless experiments to lift conversion by 0.3%.

At the same time, new transaction layers are emerging that take a much larger share.

Consider a simple example. For an average €100 online order or reservation, the costs are:

• Shopify fees: ~2.9% + €0.30
• Payment processing: ~€3.20
• AI shopping transaction fee (OpenAI fee): ~4%

Before you have paid for the product, fulfillment, or marketing, more than €7 has already disappeared into infrastructure. If your brand or store or restaurant is operating on 10–15% net margins, a 4% fee is not a small operational cost. It is often 25–40% of your margin!

For hospitality companies, add some more for reservation platforms and you can easily see that margins will become super thin when AI Commerce takes up half.

Yet the financial side is only part of the story. The deeper shift is about who owns the customer relationship.

When a purchase happens inside an AI interface, the customer does not necessarily visit your store, brand or hotel. They may never see your website. They may not join your email list. They may not follow your company at all on social media or on your newsletter list.

From the customer’s perspective, they bought something through the assistant that helped them. Not directly from you; not even aware it was you, often.

This is a familiar pattern. Distribution platforms have always traded access for control. But (AI driven) conversational commerce accelerates it significantly, and there might not be a way back. If more purchases happen through these type of intermediaries, brands, stores and restaurants risk to become ‘suppliers’ inside someone else’s interface.

The transaction remains yours. The relationship may not.

This is why many companies are starting to look more closely at owned customer programs, especially loyalty. Not simply as a discount engine. But as a relationship structure.

A well-designed loyalty program does something simple but powerful: customers choose to stay connected to the brand itself. They choose to receive your updates. They choose to participate. They choose to belong to a brand community.

That difference matters.

A customer on a list is just data. A customer who actively participates in a loyalty program has opted into a relationship. They recognize your brand, store or restaurant beyond a single transaction.

This is also where modern marketing technology can help marketing teams stay independent of external platforms. The goal is not to avoid new channels. They can be valuable sources of discovery and revenue.

But they should not be the only place where customer relationships exist.

Platforms like the VEMT help brands structure those relationships through:

• loyalty programs that reward ongoing engagement
• campaigns that capture rich first-party customer data (you own that!)
• automations that maintain communication after the first purchase
• simple workflows that marketing teams can operate themselves

The technology itself should not become another barrier. Marketing teams change roles frequently, and tools need to remain intuitive so they can be used immediately.

The real value comes from the data and relationships brands build themselves.

Because the long-term question is not whether AI shopping will grow. It will. The question is whether brands still have a direct connection with their customers when it does. If discovery increasingly happens through platforms, the brands that remain strong will be those that turn transactions into relationships. Loyalty programs are one of the most effective ways to do that. Not because they store customer data.

But because they create a reason for customers to stay connected with the brand itself. The moment to build those relationships is before they become difficult to establish.

When customers still visit your store. When they still interact directly with your brand. Once purchasing shifts fully into external interfaces, reconnecting with those customers becomes much harder.

At that point, brands are not building a customer base anymore. They are renting access to one. You can change that now.