DE / EN

BUSINESS NEWS

JENS BERGERMANN

The way consumers search for products online is on the verge of the biggest transformation since the invention of the search engine. While we've spent the last two decades getting used to typing keywords into search bars and scrolling through endless result lists, a new technology is now taking the wheel: AI Shopping Agents. These digital assistants are no longer just advisors but are increasingly acting as autonomous buyers.

For brands and manufacturers, this means a fundamental shift in priorities. It's no longer just about appealing to the human eye, but about mastering the "language of machines." Those who lay the groundwork today will secure access to the shopping carts of tomorrow. At SURS, we are closely monitoring this development, as it massively increases the demands on distribution and data management.

From Chatbot to Autonomous Agent: A New Era

Until now, we've known AI in e-commerce primarily as reactive chatbots that answered predefined questions. However, the crucial leap is from so-called Large Language Models (LLMs), which generate text, to Large Action Models (LAMs). While an LLM can explain which running shoes are suitable for a marathon, a LAM can predict and execute the next action.

Imagine just telling your smartphone: "I need new running shoes for asphalt, budget $150, size 10.5." An autonomous agent not only searches for options but also analyzes reviews, compares delivery times, checks return policies, and completes the purchase directly. These agents understand user interfaces and can interact directly with marketplaces via APIs (interfaces). The human role shifts from being an active searcher to a final approver—or even delegates the process entirely.

Pilot Projects of the Tech Giants: Who is Leading the Race?

The major platforms are investing billions to own the customer interface. Here is an overview of the current status (January 2025):

Amazon Rufus, in particular, already shows where things are headed. The assistant answers complex questions directly in the product listing, such as: "Is this rain jacket waterproof enough for heavy rain in the high mountains?" It uses current developments in AI agents and the structured data provided by retailers.

Impact on E-Commerce: When Machines Become Customers

When machines do the shopping, the rules of marketing change. We're talking about the rise of "Machine Customers." An algorithm doesn't care about an emotional lifestyle image. It doesn't scan hero images for their aesthetics; it needs structured "hard facts."

In our experience, AI crawlers today primarily weigh two areas: commercial relevance signals and the depth of product data. Crawlers pay increasing attention to the traffic of individual (sub)pages as well as the user's dwell time. In parallel, the granularity of attributes is crucial. The more precisely and comprehensively a product is described by specific attributes, the higher LLMs rate the probability of a successful conversion. AI models prefer products with excellent data quality over poorly maintained offers, as this reduces the risk of a bad purchase.

Brands must therefore make the shift from classic Search Engine Optimization (SEO) to Artificial Intelligence Optimization (AIO). An agent makes decisions based on:

• Real-time availability.

• Exact technical specifications (dimensions, materials, standards).

• Transparent pricing tiers and shipping conditions via API.

To succeed in this world, optimized product data and attributes are the most important currency. If your data isn't prepared in a machine-readable format, you simply don't exist for the shopping agent.

The Role of Marketplaces in an AI-Driven World

Why will marketplaces like Galaxus, Manor, or Amazon become even more dominant in this new era? Because they function as aggregated, trusted data sources. For a shopping agent, it is more efficient to crawl a platform with standardized interfaces than to visit thousands of individual webshops with different structures.

A key factor here is trust. Agents prioritize structured data according to Schema.org. Information like MerchantReturnPolicy and ShippingDetails is particularly critical. Without this explicit information, the agent considers the risk of a bad purchase too high. A comparison of platform algorithms shows that marketplaces offer an enormous structural advantage here. Also, AI features on Swiss marketplaces are evolving rapidly to serve exactly these interfaces.

Forecasts: What Do the Researchers Say?

According to recent studies on machine customers by Gartner, by 2030, about 15 to 20% of all company revenue will be generated directly by or significantly influenced by machines. The projected global market volume of $30 trillion makes it clear that this will not remain a niche phenomenon.

Regarding the return rate in the Swiss market, an interesting development is expected. While shopping agents simplify access and speed up the purchasing process, consumers' actual purchasing power remains stable. Precision is key: an AI agent can often match requirements more accurately than a human, which could minimize incorrect orders.

"The shopping agent will probably be more precise in its selection than a human. It's therefore quite conceivable that the return rate might even decrease as a result of this use." – Jens Bergermann (Founder of SURS)

The central question of the future will be less about returns and more about the default settings of the agents. Whether a system defaults to ordering from Amazon, Galaxus, or another provider will significantly change the retail landscape.

Conclusion: Prepare Your Business Today

Shopping agents are not a distant future—they are already in the testing phase. The success of your brand in Switzerland will depend on how "AI-ready" your data is. Those who do their homework on data quality and maintain a presence on the leading marketplaces will be found and bought by the agents.

As your partner, SURS AG supports you in positioning your brand technically and strategically so that it becomes the first choice for both humans and machines. We handle the distribution and management for you on all relevant platforms so you can profit from this technological revolution.

Frequently Asked Questions (FAQ)

What is an AI Shopping Agent?

An AI Shopping Agent is software based on artificial intelligence (often LAMs) that autonomously prepares purchase decisions for a user, compares products, or executes transactions directly.

What shopping agents already exist?

Prominent examples include Amazon Rufus (already in beta in major European markets) and Google's AI-powered search (SGE). OpenAI is also working on an agent called "Operator" that can handle general web tasks.

How do I optimize my products for AI agents?

The focus is on structured data (Schema.org), complete attributes (material, origin, exact dimensions), and providing real-time data on inventory and shipping costs. Emotional storytelling remains important for humans, but technical precision wins over the agent.

Is there a golden rule for SMBs to become "Agent-Ready"?

The most important measure is the uncompromising maintenance of data quality. This does not require an immense budget, but primarily organizational effort and structure.

"Maintaining and steadily improving excellent data quality doesn't necessarily require a large budget. It rather demands consistent structuring and organization within the company."  – Jens Bergermann (Founder of SURS)

Surs AG - Warehouse
Alte Buchserstrasse 10
8108 Dällikon
Switzerland

Surs AG - Billing
Alte Buchserstrasse 8
8108 Dällikon
Switzerland

©Surs AG 2025, Design by Surs

AT WORK