How Asakana announced its approach to food distribution efficiency on X

Asakana reads the texts, phone calls, voicemails, emails, and WhatsApp or WeChat messages that restaurants already send to food distributors, then turns them into order drafts in the ERP or spreadsheet the warehouse already uses. Unclear lines are flagged for a person before anything is entered. The company is launching from Y Combinator's Fall 2026 batch into a business where earlier ordering software failed because teams would not adopt a new app, and where thin net margins leave little room for a desk that spends the day retyping. "40 salmon for Friday" is a normal order in this trade, and it often does not say fillets or pounds. Asakana checks how that customer has ordered before, matches the shorthand to the catalog, and writes the line so a rep does not have to guess at night. The same workflow covers credit memos, returns, cutoff reminders, past-due collections, and purchase orders, including a photo of a handwritten note a driver brings back. When a salesperson leaves, the record of who means fillets and who means cases stays with the company. Yudai Matsuoka, who traded tuna at Mitsubishi, worked as a fisherman, and later consulted at McKinsey, founded Asakana with Benjamin Solis-Cohen, a former Google senior AI research engineer who worked on Gemini and computer-use agents. The company says it is already live with US seafood distributors. It is built for the product aliases, grades, pack sizes, and customer preferences that make food orders hard to read, with seafood as the strongest fit so far.