"AI is used far too inflationary – including in online retail."

Artificial intelligence (AI) is currently the buzzword that almost every industry is adorned with. From recruiting to logistics – companies promise leaps in efficiency and better customer experiences. However, if you take a closer look, it often becomes clear that many of the so-called AI applications are merely rule-based automations or statistical pattern recognition – useful, but far from true intelligence.

At first glance, AI seems to be omnipresent in e-commerce: product suggestions, analysis of surfing behavior, demand forecasts, and returns prevention – there is almost no step in the customer journey that is not supposed to be optimized, streamlined, or made more transparent by AI. Nevertheless, despite many pilot projects, the big breakthroughs have not yet materialized. This is because many systems are based on insufficient data or fail to meet customer expectations in their implementation.

AI in online shopping: Seizing opportunities

The potential is indeed huge: intelligent recommendation systems help to offer products in a more targeted manner, while size recommendations and virtual fittings can reduce buyer’s remorse. Zalando, for example, relies on AI models from Zurich that use body data to suggest the perfect fit. AI is also already being used in logistics to plan return flows more efficiently and reduce CO₂ emissions.

At its core, it is always about improving processes to win over and retain customers. Unfortunately, the "customer return experience" – i.e., how a customer experiences the return process – is still too often neglected. Yet it is precisely in this section of the customer journey that artificial intelligence can help by analyzing reasons for returns, automatically guiding customers through return processes, or speeding up service with the help of chatbots and efficient, customer-centric communication. Used correctly, this saves costs and improves customer satisfaction – and thus also sales in the medium term.

However, acceptance is limited. A recent survey by Trusted Returns shows that around 35 percent of online shoppers are undecided or find AI tools unhelpful when shopping. Many customers feel overwhelmed by automated processes or distrust decisions they cannot understand. Human advice therefore remains indispensable, especially for highly priced or emotionally charged products.

Limits of AI: Humans remain irreplaceable

AI can therefore speed up processes, but it cannot replace humans. The question is, however, whether people are even willing to accept AI as a real problem solver – especially in areas where trust and personal interaction count. Many customers prefer to communicate with people who understand them rather than with anonymous systems. AI lacks what is crucial in customer relationships: emotional intelligence, empathy, and situational sensitivity.

In addition, the database is often too weak or too limited to make informed decisions. Especially in returns management, where the reasons for triggering the returns process are manifold – from fit problems to false expectations to impulsiveness – the systems often only provide partial answers. AI can recognize patterns, but it does not understand why customers return items.

Instead of replacing humans, AI should therefore provide targeted support – for example, by taking over routine tasks, relieving employees and enabling them to focus more on emotional and empathetic customer interaction.

Another problem is that the hype surrounding "artificial intelligence" often overshadows pragmatic solutions. Many retailers have not even established simple rule-based automation – for example, for classifying reasons for returns or prioritizing customer service. Yet these systems would often be more transparent, easier to control, and more comprehensible for customers.

The issue of data protection must also be viewed critically. Personalization relies on extensive data collection – a balancing act between convenience and control. Those who want to know too much risk losing the trust of their customers.

Artificial intelligence

AI is a powerful tool, but it is not currently a panacea. Used correctly, it can support decision making processes, reduce returns, and improve customer experience. However, it should not be seen as a substitute for human expertise and soft skills, but rather as something that complements them. What is needed is KII: the intelligence to use artificial intelligence sensibly.

Online retail is at a crossroads. Some use AI strategically and responsibly and can gain real competitive advantages. While others blindly follow every trend and risk expensive investment blunders and disappointed customers.

In the end, one simple truth remains: not every solution needs artificial intelligence – sometimes a clear process is enough. True progress lies in combining technology with human understanding. Because that is what really convinces customers – not AI, but the intelligence in using it.