AI Finds the Hidden Margin

Inside the AI Transformation of Parts, People & Processes

Hi, I’m Lily. I live in the world of distribution, where margin is everything and has always been razor-thin. What's new is where the margin is hiding. It's in the returned sweater that gets back on the shelf before the season ends. It's in the tariff simulation that runs overnight and saves millions. It's in the food distributor who finally stops doing manually what a platform can do instantly. This week, the story is about margin recovery through AI adoption, and who's finding it right now.

If you only read one thing this week, it is this:

American Eagle and Dollar General are using AI to forecast demand at the ZIP code level, simulate tariff scenarios in real time, and redirect inventory, turning supply chain volatility into a competitive edge. AEO's Four-Layer AI Stack: Forecasting → Inventory Repositioning → Carrier Optimization → Network Orchestration, with the call-out stat: 60%+ tariff cost reduction as the outcome of the system working end-to-end.

What’s Working in the Field

97% of Manufacturers Call AI a Requirement

Fictiv and MISUMI's 11th Annual State of Manufacturing & Supply Chain Report, surveying 300+ senior leaders, confirms AI has crossed the threshold, becoming the primary operational standard in the industry. 97% say AI is already embedded across core manufacturing and supply chain workflows, and 95% call it a business requirement. Leaders are projecting productivity gains of 50% or more as workflows are redesigned around automation. Meanwhile, 83% of engineers still spend four-plus hours weekly on procurement tasks, what a clear target for AI efficiency gains! Tariff pressure is accelerating the shift: 99% now say supplier tariff expertise is essential in partner selection.

American Eagle Used AI Simulations When Tariffs Hit… And Made Millions

At Manifest 2026, which InstaLILY AI also attended, American Eagle Outfitters revealed a four-layer AI system that forecasts consumer demand down to the ZIP code, dynamically repositions inventory, optimizes carrier selection by cost and capacity, and orchestrates all three into a unified supply chain view. When tariffs hit in April 2025, AEO ran real-time network simulations to evaluate sourcing shifts and freight mode changes, decisions executives described as worth millions of dollars in impact. The company also expects to reduce its total tariff burden by more than 60% by early 2026.

McKinsey: The $200 Billion Problem AI Is Finally Solving in Reverse Logistics

U.S. consumers returned nearly $1 trillion in merchandise in 2024, more than double four years prior (!), costing retailers an estimated $200 billion annually, and most are still managing those returns the same way they did during COVID. McKinsey argues that AI-driven dispositioning, routing each returned item to its highest-value outcome in real time, can push value recovery from 50% to 75% by getting products back to market before demand fades. The lever is data: integrating customer history, product margin, seasonality, and demand forecasts into a single AI decision engine that acts at the moment a return is initiated, not after it enters the network. For distributors and supply chain operators, reverse logistics is no longer a back-end cost to manage. It's a margin opportunity to capture.

Pepper Raises $50M to Bring AI to the Food Distributors Still Running on Manual Workflows

Pepper has raised $50 million in funding to accelerate AI deployment across the independent food distribution market, a sector representing more than $1.4 trillion in annual sales where most operators still run on manual, fragmented systems. The platform already supports 500+ distributors and $30 billion in annual GMV, integrating with more than 70 ERP systems to bring legacy operations into a unified digital stack without costly system replacements. New funding targets AI expansion across five workflow modules covering commerce, sales, finance, payments, and customer communication. For independent distributors squeezed by consolidation from larger players, Pepper is positioning AI-powered automation as the equalizer by giving smaller operators the operational firepower of industry giants.

GE Aerospace Is Spending $300M on AI Robotics.

The International Federation of Robotics (IFR), a global robotics industry body spanning 50+ countries, projects AI-powered robotics will be standard across industries within 5-10 years, driven by faster ROI, fewer errors, and lower maintenance costs. The key shift is "Physical AI": robots that train in virtual environments and operate by experience, not programming, making them far more adaptable for logistics, warehousing, and manufacturing. For distribution and manufacturing operators, the message is: AI robotics is not a research project, but rather a capital allocation decision.

What’s In My Ears

An AI Research Primer with Kevin Reid-Morris

In this episode of the MDM Podcast, researcher and author Kevin Reid-Morris discusses his landmark two-year study cataloging over 50 real-world AI use cases specifically in wholesale distribution. The conversation opens with a striking data point: in 2023, 65% of distributors had yet to apply AI to a single business function. Today, 84% have implemented AI in at least one area. Reid-Morris offers a forward-looking take on M&A, noting that PE firms and AI-native companies are beginning to acquire legacy distributors as an entry point, a dynamic that he says would have been unthinkable two years ago.

Lily’s Quick Take

Disruption keeps arriving on schedule, and the distributors and operators who built AI into the bones of their business are the ones turning chaos into margin. American Eagle didn’t panic when tariffs hit. They simulated. Pepper isn’t waiting for independent distributors to catch up to the giants. It’s handing them the same weapons. And the IFR isn’t predicting the future of AI robotics anymore. GE Aerospace is funding it.

The lesson isn’t that AI saves the day. It’s that AI only saves the day if you trusted it enough to build it before the day arrived.

Until next week—keep your systems learning!