Amazon reduces returns with AI-powered innovation

In Germany, the number of returns has increased significantly in recent years: almost every fourth order is returned by Germans. The main reason: Clothing and shoe sizes are scaled differently from manufacturer to manufacturer. For many consumers, it is difficult to find a product with the right size online. This results in many returns due to the fit. Although returns may be very practical for consumers, it is associated with a high carbon footprint.

To counteract this dilemma, the online shipping giant Amazon has now implemented an innovative solution: artificial intelligence (AI)-supported fit advice.

The most important facts at a glance:

• Approximately one in four orders is returned in Germany, mainly due to difficulties in selecting sizes
• Amazon implemented AI-powered features including personalized size recommendations, Fit Review Highlights through customer reviews, improved size charts, and a Fit Insights tool
• Although the AI ​​initiative can help reduce the return rate and thus reduce CO2 emissions, further measures in the area of ​​sustainability are required

The four current AI functions:

The new features of the AI ​​generated by Amazon include four functions that are intended to curb returns in the long term

1. Personalized Size Recommendations:

The AI ​​and machine learning models from Amazon take into account size relationships between brands and the ratings and fit preferences of end consumers. Sizes and fits from different brands are combined through anonymous group formation. An example: If one brand's hoodie fit in an L, the AI ​​function finds out what size corresponds to that fit in another brand.

2.Fit Review Highlights:

Customers who have already purchased a garment have the opportunity to rate the product based on size accuracy, fit and material . A so-called Large Language Model (LLM) AI generates clear review highlights from this information. This means consumers no longer have to comb through reviews to find out whether the specified size corresponds to the actual size.

3. Size Charts:

The AI ​​form of Large Language Models (LLMs) is also used to make size tables even clearer and more differentiated. Unnecessary or duplicate information is filtered out and the important information is arranged better. This means that size tables should be easier to understand in the future.

4. Fit Insights Tool:

But the new AI functions are not only beneficial for consumers, they are also for the brands themselves: With more data from consumers, brands can better identify what really caused a return. This means they can continue to adapt their fit, size or material to the wishes of their customers in the future.

Reducing the flood of returns - But sustainability requires more

AI-supported advice in the form of data is exactly the right thing for an online giant like Amazon, because it has countless information from people all over the world. This means that the new form of size advice can be successful in any case. Also for all other companies from the fashion industry this form of emission saving can be a suitable method.

By minimizing returns, the online retailer can at least take a small step towards reducing CO2. Nevertheless, that's not enough: to be a real and valid sustainable company, Amazon must drastically reduce its CO2 emissions, rethink its basic returns management, improve working conditions and produce significantly less packaging waste.