AI Menu Builder
Turns a restaurant's existing menu into structured, priced catering packages, cutting a multi-hour manual task down dramatically.
AI Lab Case Study · Pre-Launch Build
Building an AI-Native Catering Marketplace
How Cruisine Was Architected to Put AI Agents Inside a Real SMB Workflow
Venture: CruisineSector: Restaurant Catering / Local CommerceStatus: Pre-launch — platform & agents built
The Challenge
Independent restaurants offering catering face high friction.
- Manual menu building — hours per restaurant to turn a menu into catering packages.
- No scalable way to answer customer questions outside phone and email.
- Multi-party payment splits that need to be accurate without manual reconciliation.
The Approach
Discover. Design. Deploy. Improve.
We reused the same four-step delivery pattern from our homepage work — but applied it to a real product build inside a live SMB workflow.
- 01
Discover
Mapped the real catering workflow, drawing on 18 years of same-day delivery operations experience.
- 02
Design
Scoped two narrow agents instead of one broad chatbot — an AI Menu Builder and an AI Concierge.
- 03
Deploy
Built a three-app architecture (customer marketplace, restaurant menu studio, shared backend) with Stripe Connect split payments and a restaurant portal.
- 04
Improve
Tested against design partners' real menu data, fixed a backend consolidation issue, ahead of public launch.
The AI Agents
Two narrow agents, not one broad chatbot.
AI Concierge
Built on the Anthropic API and grounded in each restaurant's real menu and policies, it answers customer catering questions in real time instead of phone tag.
What's Built
Status before launch.
A pre-launch snapshot of the platform.
| Component | Status |
|---|---|
| Customer marketplace | Built |
| Restaurant menu studio | Built |
| Shared backend | Built |
| AI Menu Builder | Built |
| AI Concierge | Built |
| Stripe Connect payments | Built |
| Design partners onboarded | 2 |
| Public launch | Not yet — in pre-launch testing |
Note: real adoption numbers will populate this section post-launch.
Why It Matters
Principles that keep an AI product responsible.
- Scope narrow agents before broad ones
- Build the data foundation before agent polish
- Test against real data before scaling
- Keep humans in the loop on money/exceptions
Your first move
One valuable workflow. One clear next step.
Bring us a process that is slow, repetitive, or difficult to scale. We'll help you determine what to automate, what to keep human, and how to prove the return.
Plan your AI roadmap