Why Generic ERPs Fail Fresh Food Logistics: How AI Agents Are Transforming Supply Chain Intelligence

The Supply Chain Crisis Nobody Talks About

Here’s a sobering statistic: 38% of food in the United States gets wasted somewhere between farms, supply chains, and restaurants. That’s not just an environmental problem—it’s a massive economic one. And if we could use AI technology to ship perishable goods just five days faster? That 38% could drop to 28%. Imagine the impact: we could feed entire countries with the food we currently waste.

This isn’t hyperbole. This is the reality that Sid Dixit, Chief Technology Officer at iTradeNetwork, is working to solve every single day.

iTradeNetwork is a 20+ year old company that’s quietly become the backbone of perishable supply chain logistics in North America. We’re talking about moving roughly one-third of all perishable goods across the US and Canada. Whether you grabbed strawberries at Kroger, chicken at a stadium concession stand, or coffee at an airport café—there’s a good chance iTradeNetwork’s software facilitated that transaction somewhere in the middle.

But here’s the thing: despite decades of operation, the perishable supply chain is still fundamentally broken. It’s manual. It’s inefficient. And generic enterprise resource planning (ERP) systems? They completely miss the mark.

Why Generic ERPs Fail Fresh Food

When most people think about supply chain software, they imagine the same systems that work for manufacturing or retail. But perishable goods are different. They operate under a completely different set of constraints.

You can’t ship strawberries and avocados in the same truck—they’ll spoil each other. Different commodities require different temperature controls, different transportation methods, and different handling protocols. A vanilla ERP system doesn’t understand these nuances. It treats all inventory the same way, which means it’s essentially useless for an industry where the fundamental rules are constantly changing.

This is why iTradeNetwork was created in the first place. Back around 2000, major players like McDonald’s realized they were all building custom EDI systems and ERPs to solve the same problem. They came together to support entrepreneurs who could build a specialized SaaS platform designed specifically for perishable food supply chains. The company was acquired by Grocer Technologies in 2010, but the mission remained the same: build technology that actually understands fresh food logistics.

From System of Record to System of Intelligence

When Sid joined iTradeNetwork a few years ago, the company was at an inflection point. They had a solid system of record—a place where buyers and sellers could log in and manage transactions. But Sid saw a bigger opportunity: what if they could transform that system of record into a system of intelligence?

This is where AI agents come in.

Think about what a typical buyer in the perishable supply chain does on a daily basis. Someone wants to order 100 crates of strawberries. The buyer needs to figure out:

• How much to buy? (Forecasting agent)
• What price to pay? (Pricing agent)
• Who to buy from? (Request for quote agent)
• How to negotiate the best deal? (Negotiation agent)

Currently, this is all done manually. A buyer might spend the entire day sending emails, making calls, and checking outdated pricing information. The USDA publishes pricing data, but it’s gathered through phone calls and emails—sometimes days old, sometimes 30% off, and always prone to human error.

Sid’s first MVP was elegantly simple: rebuild a real-time pricing index for perishable commodities.

The Strawberry Story: MVP Strategy in Action

Here’s where the decision-making gets interesting. iTradeNetwork handles hundreds of different commodities. Should they build pricing intelligence for all of them? Most of them? A few?

Sid’s team made a deliberate choice: focus on the commodities that mattered most to their early adopters. They reached out to customers and asked which items were most critical, most perishable, most susceptible to price swings. The answer? Strawberries and apples.

This sounds simple, but it required saying “no” to a lot of customers. It meant focusing the entire team’s energy on going deeper into just two commodities rather than spreading thin across dozens. Because here’s the thing: when you’re trading apples, one customer might call them “Apple Gala,” another might say “Fuji,” another might just say “APPL.” To create any meaningful data harmonization, you need AI agents that can scan, analyze, and understand that these three different labels represent the exact same thing.

The payoff? They took the average equipment pricing for fresh strawberries from being off by seven days (and sometimes 30%) to real-time pricing with a tiny margin of error. For the first time ever, the perishable supply chain had something equivalent to a stock ticker—a real-time index for food commodities.

Building the Roadmap: Customer Feedback and Hackathons

But the MVP wasn’t perfect. When they first launched the pricing index, they built it in pounds. Sounds reasonable, right? Except large buyers and sellers don’t think in pounds—they think in cases. One customer interaction revealed the gap, and the team had to pivot.

This is where iTradeNetwork’s product development philosophy becomes clear: continuous customer feedback loops. They run beta trials. They ask early customers for detailed feedback. They iterate based on what they learn.

More innovatively, they run in-person hackathons with customers. In Salinas, California—the heart of fresh produce country—they invite six customers and ask: “In the next six hours, can we build something to solve your core problem?” The team then takes that feedback through the entire development cycle: architecture, design, code, QA, deployment, security. All in 4-5 hours.

Some of the best ideas that come out of these hackathons are making it into their product roadmap.

The Order Agent: Bringing Small Farms Into the System

One of the most elegant solutions to emerge from this process is the “order agent.” Here’s the problem it solves: small farms don’t use modern SaaS software. They still email their produce to distributors. A typical email exchange takes 30 minutes of back-and-forth negotiation.

The order agent lets a farmer send an email, and an AI system responds asking clarifying questions: “What’s your price? Should we adjust for quality? Can you confirm this quantity?” The agent handles the conversation and automatically creates an order in the larger system. Farmers can even upload PDFs of their inventory, and the system extracts the information.

The result? Small local farms can now participate in the supply chain without needing to adopt complex software. And consumers can find local apples, bread, and fresh produce in their neighborhood stores.

The Bigger Picture: AI Agents as a Competitive Advantage

What Sid is building at iTradeNetwork represents a fundamental shift in how supply chain software works. Instead of asking “how do I do this task?” (which requires users to learn the system), the software asks “what should I do?” and handles the execution.

This is the difference between a system of record and a system of intelligence. And in an industry where 38% of food is wasted, where supply chain coordination is still largely manual, where small players are locked out of the system—that difference could be transformational.

Want to hear more about how Sid and the team at iTradeNetwork are using AI agents to transform perishable supply chain logistics? Listen to the full episode of Code Story to dive deeper into their technical decisions, scaling challenges, and vision for the future of fresh food supply chains.