S12 Bonus: The Perishable Supply Chain Crisis: Why Generic ERPs Fail Fresh Food Logistics and How AI Agents Are Transforming Error-Free Order Intake with Sid Dixit, Chief Technology Officer at iTradeNetwork

Episode Summary

Executive Summary: Sid Dixit, CTO of iTrade Network, explains how the company is transforming perishable supply chains from a manual system of record into an AI-driven system of intelligence and autonomous agents. He details launching a real-time cost index, using customer feedback to refine products, hiring and mentoring AI-native graduates, scaling AI with evals/observability, and his broader mission to democratize AI building while reducing waste and improving supply chain efficiency.

Main Topics: iTrade Network’s role in perishable supply chains (Priority: 5/5): Sid describes iTrade as a SaaS backbone for roughly one-third of perishable supply chain activity in North America, helping buyers and sellers handle orders, invoices, pricing, and negotiations for goods like strawberries, apples, chicken, and cafeteria food. From system of record to system of intelligence to agents (Priority: 5/5): He frames the company’s AI evolution as moving beyond logging transactions to recommending actions and automating workflows, such as forecasting, pricing, RFQs, and negotiation agents. Building Cost Index as the first AI MVP (Priority: 5/5): The team created a real-time pricing index from anonymized transaction data to replace stale USDA pricing, giving customers near-stock-market-like price guidance for perishables. Customer-driven product decisions and roadmap (Priority: 4/5): Sid emphasizes beta trials, continuous feedback, and a where-to-play/how-to-play/right-to-win framework to choose commodities, adjust unit logic, and prioritize new products like Order Agent. Scaling AI systems reliably (Priority: 4/5): He discusses the challenge of running AI agents over millions of daily transactions, including model drift, the need for evals/LLM judges, and observability tooling to maintain accuracy and trust. Talent, culture, and leadership principles (Priority: 4/5): Sid values builders, hires fresh graduates, and reinforces leadership principles like ownership, bias for action, customer obsession, trust, and curiosity as the basis of iTrade’s culture. Mission to close the AI divide (Priority: 4/5): Beyond company products, Sid is building 1million.build, a free program to teach one million people how to build with AI and avoid being left behind by automation.

Key Arguments: Perishable supply chains are uniquely complex and need domain-specific software because standard ERP/EDI approaches don’t handle spoilage, mixed loads, or commodity nuances well. AI can turn supply chain software from a passive system of record into a proactive system of intelligence that recommends and automates decisions. Real-time, transaction-derived pricing data is superior to stale, self-reported commodity pricing and can materially improve purchasing decisions. Customer feedback should determine product priorities; starting with high-value, high-variance commodities like strawberries and apples created the strongest initial value. Small farmers and other low-tech suppliers need lightweight entry points like email and PDF ingestion to participate in modern supply chains. At scale, AI quality depends on rigorous evals, classification harmonization, and observability; without them, accuracy degrades as volume increases. Hiring and mentoring AI-native graduates is both a workforce strategy and a way to build a future-ready engineering organization. Teaching people how to build with AI is a social imperative because automation will widen the gap between those who can use AI and those who cannot.

Data Points: Food wasted in the U.S.: 38% - Sid cites national food waste as a major supply chain problem AI could help reduce. Food waste reduction potential: 10% less - He says shipping food five days faster could reduce waste by about 10%. iTrade Network market coverage: Around one-third - He says the company moves maybe one-third of perishable supply chain activity in North America. Company age: Almost 27 years - Sid notes iTrade has been operating for nearly 27 years. Early commodity focus: 30 commodities - The team narrowed the initial cost index scope to 30 commodities. First launched commodities: Strawberries and apples - He says these were the first two commodities for Cost Index. Pricing freshness improvement: From 7 days to 2 hours - He contrasts USDA pricing lag with iTrade’s near-real-time index. Pricing error reduction: Off by 20-30% to very small margin of error - He says traditional pricing can be significantly inaccurate compared with the new index. Transaction scale: 1 million transactions per day - Sid uses this scale to explain AI harmonization and classification challenges. Model quality improvement: 90% / 85% to 98-99% - He says evals and LLM judges improved accuracy, precision, recall, and F-scores to near 99%. Feedback event: 6 customers - He describes a hackathon in Salinas, California with six customers participating. Hackathon build window: 4 to 5 hours - Teams took ideas from concept to deployment quickly using AI agents. AI education target: 1 million people - 1million.build aims to teach a million people to build with AI. Course duration: 18 days - He says someone with no coding background could learn the full build/test/deploy workflow in 18 days. Hiring plan: 10 to 12 open positions - Sid says he is hiring fresh college graduates for these roles. Microsoft hardware issue: 220-volt chargers - He recounts the charger recall story involving Europe-bound devices.

Pivotal Quotes: "“We can provide not only buyers and sellers how to do this by using your hands, keyboard, and mouse, but more importantly, maybe what to do.”" — Sid Dixit: He explains the shift from a system of record to a system of intelligence. "“If we can use this AI technology to ship it even five days faster, that number could be 10% less.”" — Sid Dixit: He links AI-enabled supply chain speed to food-waste reduction. "“Do not build anything which is not unmet customer need or customer pain point.”" — Sid Dixit: He offers advice to young entrepreneurs on product focus and validation.

Implications: The episode shows AI is moving into practical, high-impact industrial workflows, not just software demos. For supply chain leaders, the competitive edge will come from real-time data, agentic automation, and customer-led iteration. For workers, AI literacy is becoming essential.