Restaurant Reset
The Owners Who Keep Their Staff Are Growing Faster | Christian Wiens, Loman
A seafood chain's menu looked completely normal online. Then Loman pulled it out of the point of sale and it broke their ingestion tool. Every item had a modifier. Every modifier had a sub-modifier of every other modifier, ten levels deep. Somebody built that on purpose, years ago, and nobody has touched it since. Christian Wiens is founder and CEO of Loman , an AI phone agent that has now handled millions of calls across roughly 1,500 restaurants. Before this he worked in cybersecurity, which is where he learned what a false positive costs you and why he would rather lose a deal than oversell what the product can do. He tells Andy Grindstaff exactly what breaks these systems, and it is almost never the AI. It is the data. A customer asks for a roast beef three way, a thing they have ordered that way for twenty years, and it appears nowhere in the point of sale. Nested modifiers nobody can explain. Middleware adding a hop that turns one error into a four-vendor investigation. Christian's team built a proprietary cleanup module rather than asking owners to fix their own menus, and he will personally spend an hour rebuilding one because he counts it as customer acquisition cost. The back half gets into where this is all going. Agents calling agents. Why voice is not going away as long as humans are humans. And the question Andy keeps putting to operators, which is whether the technology should replace a person or free the people already on the floor to do the part a machine cannot. In this episode: • The seafood menu that broke the system, and what was actually inside it • The roast beef three way problem, and why the AI had no chance • What had to be built around the model before it could be trusted on a real call • Building trust with the owner, and separately with the caller • What cybersecurity taught him about false positives and overpromising • Where the chain really fails, and why middleware is usually the culprit • Why Loman cleans the menu instead of asking the restaurant to • Build-your-own-bowl concepts, and why they are the hardest • Why he will spend an hour fixing a menu himself and call it CAC • Agents booking with agents, and what that does to the phone • How an operator should actually measure whether the technology worked Subscribe for more conversations with the operators actually running the restaurant. Restaurant Reset is available on YouTube, Spotify, and Apple Podcasts. Restaurant Reset is brought to you by Genius™ (Link globalpayments.com/genius) from Global Payments, the restaurant point of sale system that can power orders, payments and services in any type of food and beverage setting. Timestamps: 00:00 Intro 01:14 Meet Christian Wiens and Loman 02:19 What had to be built around the model before it could take a call 07:11 The roast beef three way that exists nowhere in the data 09:48 Building trust with the owner and with the caller 15:50 Why the phone was the lowest hanging fruit 16:50 What cybersecurity taught him about false positives 20:18 Where the tech chain actually breaks 21:55 Middleware, and why they push for direct integrations 23:50 Menu structure as the big gotcha 24:52 Why Loman cleans the data instead of the restaurant 26:35 Build-your-own-bowl concepts 29:07 “The menu looks fine” 29:49 Spending an hour rebuilding a customer's menu 31:19 Agents calling agents 33:50 Why voice is not going away 39:27 AI should make hospitality more human 40:08 How an operator should measure whether it worked 44:35 The moment he knew they were onto something 45:16 Where to find Loman
