AI is overtaking your own website as the shopper's preferred route

In a Bloomreach survey of more than 4,000 shoppers in the U.S. and the U.K., 41.4% said that, if they had to choose, they would prefer to shop through an AI assistant rather than directly through a brand’s website. The latter option is still preferred by 38%.

A year earlier, in a separate Bloomreach survey, 58.9% still preferred the direct route to the brand’s website.

These figures are from the U.S. and the U.K.; they are not a Belgian benchmark. And since these are separate studies, you should be cautious about making a direct comparison.

But the trend is hard to ignore: an increasing portion of the customer journey is shifting to an AI layer between the customer and the brand.

Shoppers no longer necessarily start their search on your homepage, in your online store, or even on Google, as we’ve been accustomed to for years. An AI assistant can search for, compare, summarize, and recommend products before the customer has ever visited your own platform.

And that changes more than just search.

The first thing that stands out, of course, is your visibility.

In 2026, Google published new guidelines on how content can perform within generative search experiences such as AI Overviews and AI Mode. The message is strikingly straightforward: for Google, the fundamentals of good SEO remain unchanged.

High-quality content. Technical accessibility. Authority. Consistent and well-structured information.

Anyone who is barely visible in traditional search results today won't suddenly become dominant in generative answers on their own.

For Google, GEO—Generative Engine Optimization—is therefore less revolutionary than is sometimes suggested. It largely builds on existing SEO principles, but raises the bar for quality, structure, and clarity.

Outside of Google, the playing field is indeed getting broader.

AI assistants can aggregate information from websites, product feeds, marketplaces, reviews, data sources, and other commercial ecosystems. As a result, discoverability is becoming less and less just a matter of optimizing your own website.

Your brand, products, and services must also be able to be found, understood, and compared accurately outside your own digital environment.

But that's only one side of the story.

In fact, there is a second trend. And it touches on the relationship between brand and customer in a much more fundamental way.

As discovery, comparison, and selection increasingly take place within an AI assistant, your brand loses some of the context and control it used to have within its own channels.

You know less about the path someone has taken before they end up with you.

A larger part of that context may lie outside your own environment.

Because the more discovery and selection take place outside your own channels, the less you can rely on the context of a single website visit, a single campaign, or a single channel.

You need to be able to reconstruct that context yourself across different channels.

That is why first-party data and an integrated 360° view of the customer are becoming more important, not less.

Not because data in and of itself is a competitive advantage, but because it allows you to identify who you’re dealing with, what that person has done in the past, what relationship they already have with your brand, and what’s likely relevant to them right now.

And that is precisely where a crucial difference arises as AI begins to support increasingly autonomous decision-making.

An AI application can only reason as well as the context you provide it.

When the same person appears as different profiles in the CRM, online store, point-of-sale data, and marketing platform, the AI does not recognize them as a single customer.

She's thinking about three half-customers.

That is also why many AI projects ultimately don't run into trouble with the algorithm itself, but with the foundation beneath it.

Unleash automated campaigns on fragmented customer data, and you’ll do more than just automate your marketing.

You're also automating your mistakes.

Faster. On a larger scale. And often without anyone noticing them right away.

That doesn't mean organizations have to spend years perfecting their data before they can start using AI.

The fundamentals and their application must evolve together.

One reinforces the other.

By the way, you don't have to wait for agentic commerce to happen.

The value of a unified customer view is already measurable today.

For example, at a Belgian fashion retailer, we identified underperforming top customers based on a single, integrated customer view. We then engaged them with short, well-timed promotions.

Without a new channel.
Without an additional media budget.

The result:

The technology itself was not the goal.

The difference arose because customer data from various sources was consolidated, translated into a usable customer profile, and then converted into a concrete action.

Agentic search and agentic commerce will likely take over more and more stages of the customer journey in the coming years.

That doesn't mean brands are becoming irrelevant.

However, it is true that a larger portion of their interaction with customers may take place in environments over which they have no control.

That is exactly why it’s becoming more important to make the most of what you do have: your own customer data, your customer relationships, and the ability to act on them intelligently.

AI doesn't make bad data any smarter. Above all, it makes the consequences of that data more quickly apparent.

Companies that have a clear understanding of their customers can therefore do more than just implement AI more quickly.

Above all, they can make better use of it.