Stockline Is Replacing the Fax Machine That Runs America's Food Supply Chain

The Macro: The $1.5 Trillion Industry That Runs on WhatsApp and Prayer

Food wholesale is one of those industries where the gap between how important it is and how modern its technology is borders on absurd. Every restaurant, every grocery store, every cafeteria in the country gets its food from wholesalers. The U.S. food distribution market is worth roughly $1.5 trillion annually. It is massive, essential, and running on software infrastructure that would make a 2005 IT manager wince.

Here is what a typical day looks like for a food wholesaler in 2026. Orders come in through WhatsApp messages, SMS texts, phone calls, and emails. Sometimes they come in through all four channels from the same customer in the same hour. A sales rep or office worker manually reads each order and types it into an ERP system. The dominant ERPs in this space are BlueLink, Produce Pro, and Edible Software. Some operators are still using custom Access databases or, in cases I wish were jokes, handwritten ledger sheets.

The manual entry process is slow, error-prone, and expensive. A mid-size wholesaler might process 200 to 500 orders per day. Each order has multiple line items with product names, quantities, units, and pricing that can vary by customer. One typo can send 50 cases of romaine lettuce to a restaurant that ordered 5. The wholesaler eats the cost on every mistake.

The reason this has not been fixed is that the problem is genuinely hard. Orders arrive in natural language across multiple channels. Customers use shorthand, abbreviations, and inconsistent product names. “10 cs roma” means 10 cases of roma tomatoes, but only if you know this customer’s vocabulary and this wholesaler’s product catalog. Traditional software cannot parse this without extensive manual configuration that breaks every time a customer changes how they order.

The Micro: A Fish Wholesaler’s Daughter and a Stanford Dropout Build an AI ERP

Stockline was founded by Mariem Ould Ismail and Moncef Djafri. The founder-market fit here is personal and direct. Mariem’s father is a fish wholesaler. She grew up watching him take orders by phone, write them on paper, and key them into outdated software. She holds a Master’s in Applied Mathematics from Ecole Polytechnique, one of France’s most competitive engineering schools, and dropped out of MIT to build Stockline. Moncef is also an Ecole Polytechnique alumnus, started coding at 13, and dropped out of Stanford to co-found the company.

They came through Y Combinator’s Summer 2025 batch as a two-person team in San Francisco. The product attacks the order entry problem head on. Stockline connects to every channel a wholesaler uses to receive orders. WhatsApp, SMS, email, phone. It pulls all incoming orders into a single unified inbox and then uses AI agents to parse, interpret, and enter those orders automatically.

The AI piece is the hard part and the part that matters. When a restaurant texts “20 cs chicken breast boneless 10lb, 15 bg yellow onion 50lb, 5 cs avocado 48ct” to a wholesaler’s phone number, Stockline’s agent has to match each item against the wholesaler’s specific product catalog, resolve ambiguities (does “chicken breast boneless” match SKU 4412 or SKU 4415?), apply the correct customer-specific pricing, and format the order for the ERP system. This is exactly the kind of messy, context-dependent language understanding problem that modern LLMs are good at, if you build the right scaffolding around them.

The approach is to replace the legacy ERP entirely over time, not just sit on top of it. Stockline positions itself as an AI-native ERP, meaning the AI is not a feature bolted onto existing software. The entire system is designed around automated data capture and intelligent processing from the ground up.

The competitive landscape in food wholesale software is fragmented and sleepy. BlueLink has been around since the 1990s. Produce Pro targets produce distributors specifically. Silo, a more recent entrant, raised $132 million and offers a broader food supply chain platform, but its AI capabilities for order entry are limited. Choco, based in Berlin, built a messaging platform for food ordering but has struggled with U.S. market adoption. None of these competitors have built an AI-first order entry system from scratch.

The Verdict

I think Stockline has one of the most tangible value propositions I have covered this quarter. If you can take a wholesaler processing 300 orders per day with three data entry employees and reduce that to one person reviewing AI-generated entries, the ROI calculation is immediate and obvious. This is not a “change how you think about your business” sale. This is a “save $120K in labor costs this year” sale.

The founder story is strong. Building enterprise software for an industry because you literally grew up in that industry is the kind of origin story that resonates with buyers. When Mariem walks into a wholesaler’s office, she is not a tech person trying to understand their world. She is someone from their world who happens to build technology.

In 30 days I want to see order accuracy rates. The AI agent needs to be right at least 95% of the time for wholesalers to trust it. Anything below that, and the review burden on the human operator negates the time savings. If they are hitting 98% or higher, this product sells itself.

In 60 days the question is channel coverage. WhatsApp and SMS parsing is one thing. Phone call transcription and order extraction is significantly harder. If Stockline can handle voice orders with the same accuracy as text orders, the addressable market doubles.

In 90 days I want to understand expansion beyond order entry. The ERP replacement thesis is ambitious. Order entry is the wedge, but inventory management, invoicing, route planning, and supplier procurement are all pieces of the food wholesale workflow that need modernization. The roadmap from “AI order entry tool” to “complete AI-native ERP” is long, and the sequencing matters.

Two Polytechnique dropouts building an ERP for fish wholesalers is exactly the kind of improbable, founder-driven vertical play that produces real companies. The market is big, the pain is acute, and the timing is right.