Every few months, it’s worth scrolling through the Shopify app store to see what new returns-management options e-commerce merchants are offering. They’re endless — so many that three screens don’t even cover them all. Yet logistics-focused solutions are scarce.

This highlights a bigger problem: For the past decade, we’ve thrown software at what is ultimately a physical issue. Returns involve real products and logistics, and as much as artificial intelligence promises to streamline things, it needs one crucial component: physical data.

The truth is that the industry has focused so much on pixels — software, predictive analytics, machine learning — that it has ignored the hard reality of parcels.

The logistics industry is eager to see AI transform returns, but the technology can’t fix what it can’t see. The biggest challenge is bridging the gap between digital data and physical goods. AI needs real-time, physical insights into returned items to work effectively.

A recent study showed that 40% of executives doubt their data is ready for AI to deliver accurate results, and reverse logistics is no exception. The returns process is complex, involving defective items, cross-border challenges and messy data. Retailers are dealing with fragmented systems, regulatory hurdles and inflation pressures. AI can’t work with incomplete data, and cleaning it up is rarely a priority. Much of the hype around AI in returns is about making tech sound cutting-edge, not solving the real problem.

Even with AI’s role in reducing impact, you can only reduce returns so much — consumers will always want the option. The real challenge is minimizing the impact of returns on consumer convenience, merchant profitability and environmental emissions. By 2024, returns are projected to exceed $550 billion, with the environmental toll rising. AI can help by automating processes and predicting patterns, but it won’t eliminate returns. Instead, we should focus on making returns less painful for consumers, cheaper for merchants, and less harmful to the planet.

AI is already disrupting reverse logistics, but in smaller ways. Some merchants are shifting from branded return portals to chat-based systems integrated with customer support, simplifying the consumer experience. AI-driven fraud detection is also emerging, using data patterns to flag suspicious returns before they hit the pipeline.

But the real frontier isn’t just automating returns — it’s integrating digital tools with the physical process. Systems that combine software and logistics are leading the way.

In the long run, AI’s greatest potential in returns lies in building customer loyalty. Generative AI could transform returns by suggesting personalized alternatives, such as like different sizes or related products, based on the customer’s preferences and purchase history. As a result, returns become tailored exchanges, and customers stay engaged.

Instant refunds are another major opportunity. AI could analyze customer-submitted images or videos to verify a return’s condition and process refunds immediately, cutting wait times. AI could also optimize return policies, rewarding loyal customers with perks like extended windows or free shipping, while nudging frequent returners toward exchanges or store credit. On top of that, AI could promote sustainability by offering eco-friendly e-commerce options to help retailers maximize value from returned goods.

So is AI ready to fix retail returns? Not quite yet, but it’s coming. Currently, it can reduce the impact of returns by improving efficiency and personalizing the experience, but until we bridge the gap between digital and physical, its potential will be limited. Retailers need AI combined with systems that physically validate returns to make everything work for consumers, businesses and the environment. Only then will we see AI’s true impact on the returns problem.

Sylvia Ng is chief executive officer of ReturnBear.