The opinions expressed here by Trellis expert contributors are their own, not those of Trellis.
The apparel industry is finally using AI in ways that matter. Not everywhere, and not always well, but enough to say the shift is real. The early gains aren’t coming from some fantasy of a robot-run sewing floor. They’re showing up in planning, forecasting, fit tools, pricing, search, digital content and internal workflow systems. That may sound mundane, but it isn’t. Those are the very places where this business has long made some of its worst mistakes.
Apparel has a habit of treating bad decisions as standard operating procedure. Brands buy the wrong colors, the wrong size mix, or the wrong seasonal story. Retailers chase volume and then spend months clawing back margin through markdowns. Suppliers absorb late changes from customers who were never as certain about demand as they claimed to be. Then, once the waste has piled up, the industry talks about agility, complexity and market pressure as if none of this could’ve been avoided.
That’s where AI has found an opening. When it works, it helps companies make fewer bad calls before the damage spreads through the chain. It can sharpen forecasts, improve allocation, support better recommendations and help customers find the right size before they buy. This leads to less excess fabric, less dead inventory and fewer returns. Multiply that across thousands of SKUs and the waste becomes structural.
The best AI uses
So far, the best uses of AI in apparel are practical ones: forecasting, fit and internal productivity. None of this is glamorous, but that’s exactly the point. AI can work through sales data, customer behavior and inventory patterns faster than most planning teams. It can improve search and product discovery. It can help employees find information that used to sit in separate systems, hidden in spreadsheets or trapped in email chains. Fit deserves special attention because returns remain one of the dirtiest habits in the apparel business. A return isn’t just a disappointed customer; it’s another trip through freight, sorting, handling, repackaging and often markdown. Any tool that helps more people get the right size the first time has real value. It cuts cost and it cuts waste.
The same goes for demand planning. Better forecasts won’t eliminate risk and they won’t turn fashion into an exact science. But even modest improvement matters in an industry where chronic overbuying has become routine. If AI helps companies buy closer to what they can actually sell, that’s useful. If it helps them avoid another pile of goods headed for discount channels, even better.
The harder question is what gets lost when speed becomes the main objective.Apparel has a long record of pushing cost pressure and risk outward, usually onto the people with the least leverage. There’s no reason to assume AI will break that pattern on its own. It could just as easily sharpen it. A faster process isn’t automatically a better one. It depends on what’s being accelerated and who’s carrying the downside.
AI’s role in manufacturing
AI is also entering manufacturing, but the reality is less dramatic than the conference circuit likes to suggest. In most cases, it’s being used for monitoring, diagnostics, maintenance support, output tracking and productivity analysis. Those applications may improve efficiency and in some factories they probably already are. But efficiency alone doesn’t settle much. Does the technology reduce waste? Does it improve planning and product flow? Does it make work more stable, or does it simply tighten control? Who gets the gain when productivity improves? Those questions matter more than the sales pitch.
The bottom line
AI is becoming a useful operating tool in apparel, but it has a way to go. The technology works best where the task is repetitive and data-heavy. That gives the industry a real opportunity. But apparel’s deeper problems are still managerial. The business still rewards speed over judgment, volume over discipline and growth over restraint. AI won’t fix that by itself.
That’s the real test now. If AI helps the industry make less of what it can’t sell, fit customers better, and reduce waste across the chain, then it will have earned its place. If it simply helps companies move faster through the same cycle of overproduction, markdowns and churn, then the tools will be new, and the outcome will be the same.