By Somdutta Singh
E-commerce is often framed as a game of ads and apps. The real challenge, however, is behind the scenes. As storefronts start to look alike, the gap is opening in how well companies run their operations.
Amazon recently announced plans to invest over $35 billion in India by 2030. The focus was on AI, logistics, and backend capacity. When big companies invest like this, they are betting on where advantage compounds.
That advantage is becoming operational. Warehouses are changing fast. Automation is moving past sorting and packing into harder tasks like truck unloading. AI decides where goods move, when they move, and how fast. These decisions ripple through every order.
Routing is evolving just as quickly. AI now selects delivery paths by weighing distance, traffic, capacity, and failure risk. Delays are avoided before they happen, reducing missed deliveries and failed attempts that damage cost and trust.
Returns make this shift impossible to ignore. Total retail returns are projected at $890 billion in 2024 and $849.9 billion in 2025. Even small improvements here matter. A one percent change moves billions.
Fraud and abuse sit inside this problem. Losses from return fraud touched $103 billion in 2024, with roughly 15% of returns flagged as fraudulent. Retailers respond by tightening rules, and honest customers often feel the fallout.
Cash on Delivery (COD) and Return to Origin (RTO) make this even harder. When a COD order fails, the product travels back to origin. Shipping costs double and the sale is often lost. At scale, this quietly eats margins. This is why AI and ML are now being used to predict RTO risk, flag risky orders early, optimise routing, and decide when COD should be offered at all.
At the same time, 76% of shoppers say free returns influence where they buy. Retailers cannot cut costs blindly. They need decisions that protect margins and customer trust together.
This is where the shift becomes clear. Data is no longer valued only for cutting costs or speeding things up. Companies are looking for data that helps them make better decisions and see what actually changed. The demand is moving away from tools that promise efficiency toward technology that adds value and delivers measurable outcomes.
But the catch is, AI depends on clean data; without it, automation spreads errors, skews decisions, and clouds performance measurement.
This is why backend excellence matters now. When the backend is strong, AI becomes a multiplier. The real edge will belong to teams that choose effectiveness over efficiency, again and again. They will track outcomes, fix root causes, and focus on what truly improved. The best companies won’t look exciting from the outside. They’ll simply work better, every day.
The author is founder & CEO, Assiduus Global