On this episode of The Restaurant AI Podcast, ClearCOGS co-founder and CEO Matt Wampler sits down with Zhong Xu, the visionary founder behind Deliverect, now valued at over $1.4 billion. Zhong opens up about the relentless execution and radical simplicity that powered Deliverect’s rise from a scrappy MVP to a global restaurant infrastructure platform used by thousands of brands.

Zhong shares the unbelievable early story: validating the business before writing real code, onboarding 300 restaurants across multiple countries within months, and using this traction to fuel what would become one of the biggest success stories in restaurant technology. From there, he and Matt dive into the complexity of the modern digital restaurant stack and why Zhong believes AI agents and the Model Context Protocol (MCP) will change everything.

They break down how MCP gives AI the context it needs to reason like a true operator, how agents will place orders, test systems, optimize menus, and even catch fraud in real time. Zhong also details the next wave coming to restaurants: drone delivery, autonomous logistics, AI-generated menus, refund-fraud prevention using camera intelligence, and operational automations that happen long before a human ever touches the POS.

Packed with insights on execution, growth, AI strategy, and the future of restaurant tech, this is the definitive episode for operators, founders, and anyone watching the rapid acceleration of restaurant AI. Zhong’s billion-dollar journey is a masterclass — and his vision for what’s coming next is even bigger.

Episode Links -

00:00 Intro

01:36 Zhong’s entrepreneurial background

11:10 The catalyst for starting Deliverect

17:00 Why traditional Point of Sale (POS) systems couldn't solve the digital ordering problem 

20:53 Fragmented POS market

28:55 The "Wizard of Oz" Minimum Viable Product (MVP) for Deliverect

31:55 Defining a working technology MVP in the restaurant industry 

38:15 State of the relationship between third-party delivery and restaurants

41:21 The MCP is the "smart layer on top of the plumbing" 

47:10 The challenge of large language models (LLMs) and the need for a protocol like MCP to prevent "hallucinations" 

57:26 Building a 'firewall' to manage which AI agents can access a restaurant's private data.

1:01:00 The specifics of what Deliverect is currently building

1:07:15 Drone delivery isn’t far away

1:12:35 How technology can improve the poor economics of the restaurant industry 

1:14:55 Final discussion on data visibility and how information should be shared across the different levels of restaurant management 

1:18:56 Outro