TapWait AI: Smarter Restaurant Operations and Marketing with Open-Weight AI

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

Every restaurant has two important jobs:

delivering excellent customer service and attracting customers. But managing service requests, coordinating waitstaff, and creating engaging promotions can be challenging when these activities rely on disconnected tools and manual processes.

TapWait AI brings restaurant operations and AI-powered marketing together in one connected hospitality platform.

Built for restaurants, bars, cafés, and food service businesses, TapWait AI combines a Flutter mobile application for waitstaff, a web dashboard for managers, real-time backend services, and AI-assisted promotional campaigns.

Here is how it works.

A Mobile App for Restaurant Waitstaff

Waiters use the Flutter mobile application to receive and manage customer service requests. Requests move through a defined workflow, allowing staff to track their status and respond to customer needs.

Socket.IO supports real-time communication, while Firebase Cloud Messaging supports push notifications, including background notifications where supported.

The goal is to help restaurant teams coordinate service more effectively and reduce the friction between receiving a request and responding to it.

A Web Dashboard for Restaurant Managers

The web dashboard gives restaurant managers a central place to oversee service requests, monitor their progress, and coordinate restaurant operations.

Instead of depending entirely on verbal communication, managers can use a connected digital workflow to stay informed about service activity.

The dashboard also provides the foundation for managing AI-generated promotions and marketing campaigns within the same platform.

MenuBoost AI: Turn Restaurant Menus into Marketing Opportunities

A restaurant's menu is more than a list of dishes. Every menu item and special offer can become an opportunity to attract customers.

MenuBoost AI uses Google's hosted Gemini API to generate promotional content based on saved menu items, prices, selected offers, marketing tone, and target platforms.

Restaurant managers can create content for channels such as WhatsApp, Facebook, and Instagram without having to start from a blank page.

By grounding generated content in restaurant-specific information, MenuBoost AI aims to make promotional creation more relevant and practical for businesses without dedicated marketing teams.

WhatsApp Marketing and Customer Engagement

TapWait AI incorporates WhatsApp Cloud API workflows to support restaurant messaging and promotional campaign management.

The vision is to help businesses turn menu updates and special offers into engaging customer communications while managing campaign workflows through the platform.

Promotional delivery must respect subscriber consent, applicable messaging policies, and WhatsApp's requirements. The platform is designed to support responsible customer engagement rather than indiscriminate messaging.

Campaign Guard: Combining Gemma with Practical Validation

Generating attractive promotional content is only part of the challenge. Restaurants also need to ensure that promotions reflect their actual menus and offers.

Campaign Guard explores using Gemma 3 270M through Ollama to review saved promotional content against its associated menu and offer information.

The workflow retrieves the saved promotion, supplies its content and source information to the local model, and returns a structured review containing a summary, potential issues, and suggested corrections.

For stronger reliability, the broader validation approach combines model-assisted review with deterministic application logic. Prices, discounts, and offer conditions should be checked against trusted data and explicit business rules rather than relying on a language model alone.

This creates an opportunity to explore how lightweight open-weight AI models can support practical quality checks within an existing business application.

Demo

Step 1: Customer Calls a Waiter


Step 2: The Request Reaches the Restaurant's waiters

Step 3: The Waiter Receives and Accepts the Request

Step 4: The Customer Receives Assistance

Step 5: The Manager Monitors Operations

Step 6: AI-Powered Marketing with MenuBoost

Experience TapWait AI: A connected hospitality platform bringing customer service, waiter coordination, management visibility, and AI-powered marketing together.

Code

Mobile App : https://github.com/Y-elv/tapwait-waiter
BAckend : https://github.com/Y-elv/tapwait-waiter-api

How I Built It

• I built TapWait AI as a full-stack hospitality platform connecting mobile, web, backend, and AI workflows.
• Mobile: Flutter provides the waiter-facing application, enabling staff to receive and manage service requests on Android devices.
• Backend: Node.js and Express power the REST APIs and application workflows. MongoDB stores service requests and promotion records.
• Real-time communication: Socket.IO supports live updates between connected application components.
• Notifications: Firebase Cloud Messaging supports push notifications for relevant waiter workflows.
• Web dashboard: The management interface gives restaurant managers a central place to monitor operations and work with promotional content.
• AI marketing: Google's hosted Gemini API powers MenuBoost AI's promotional content generation.
• Local AI experimentation: Gemma 3 270M, run through Ollama, powers the Campaign Guard experiment for reviewing saved promotional content against its source information.
• Messaging integration: WhatsApp Cloud API workflows support restaurant messaging and promotional campaign management.
• One of the main engineering challenges is making these components work together as a coherent product rather than a collection of independent features. The goal is to connect service operations and marketing through consistent APIs, saved records, clear workflows, and a practical management experience.

The platform is evolving through integration testing, workflow improvements, and feedback from potential users. My focus is on making its core features dependable and useful in real hospitality environments.

Why Does Open Innovation Matter?

Open innovation gives developers the flexibility to explore different AI models, combine approaches, and choose suitable deployment strategies for specific tasks.

For TapWait AI, this means exploring a hybrid AI architecture:

• Gemini for content generation: A hosted model generates promotional content using restaurant menu items and selected offers.
• Gemma for local review: Campaign Guard explores local inference using Gemma 3 270M through Ollama to review promotional text against supplied source information.
• Deterministic validation: Application code can verify prices, discounts, and other explicit business rules against trusted records.
• Flexible architecture: Separating generation, review, and business-rule validation allows each component to evolve independently.

This is particularly interesting for hospitality businesses because AI should be useful, adaptable, and grounded in actual business information. A local model can also provide an alternative inference path when the application is configured to run it locally, although its suitability depends on hardware, latency, and review quality.

The wider opportunity is to make AI-assisted business workflows accessible to organizations that need practical tools rather than complex, disconnected systems.

Who Would Pay for This?

TapWait AI is designed for:

• Independent restaurants looking to improve service coordination.
• Bars and cafés managing busy customer-facing operations.
• Small food businesses that need a simpler way to prepare promotional content.
• Hospitality groups interested in connecting operational workflows with marketing tools.

The business opportunity is to offer a connected platform that helps restaurant teams manage service requests, coordinate staff, prepare promotions, and organize customer engagement without juggling multiple disconnected applications.

Potential business models include monthly subscriptions, tiered plans based on business needs, and higher-tier offerings for multi-location operators and additional integrations.

The focus is on creating measurable value through more organized workflows, easier campaign preparation, and better visibility into restaurant operations, rather than promising unverified revenue increases.

My First User

A neighboring startup has expressed interest in using TapWait AI and has been encouraging about the idea.

That response has motivated me to keep improving the platform and focus on practical business needs rather than building features without a clear purpose.

My next step is to move toward structured real-world use, gather feedback, observe how the workflows perform, and identify where TapWait can save time or make restaurant service and marketing easier.

This feedback will help guide product improvements and validate the platform's value in an actual business setting.

Prize Categories

Best Use of Gemma

TapWait AI explores using Gemma 3 270M through Ollama for Campaign Guard, a local AI review workflow within a broader restaurant marketing platform.

The feature retrieves a saved promotion and compares its generated content with the associated menu and offer information. Gemma produces a structured review intended to highlight potential inconsistencies and suggest corrections.

The approach combines open-weight AI experimentation with application-level validation, recognizing that prices, discounts, and explicit offer rules require reliable checks against trusted data.

The broader idea is to demonstrate how a lightweight local model can complement a hosted generation model within a practical business workflow.

TapWait AI's vision is simple: connect restaurant service, intelligent marketing, and practical AI in one platform built for the needs of modern hospitality.