Makondo AI Manager

An AI restaurant manager that turns sales, inventory, recipes and costs into daily recommendations for purchasing, waste reduction, pricing and profit.

Comment

Inspiration

Makondo AI Manager was born from a real problem.

I own a Colombian restaurant called Makondo, and every day I had to check inventory, employee hours, recipes, purchases and finances manually.

Instead of using spreadsheets and multiple disconnected tools, I wanted a single AI assistant capable of understanding everything happening inside the restaurant.

The goal was simple:

Ask questions in natural language and receive useful answers based on real business data.

What it does

Makondo AI Manager is an AI-powered restaurant operating system.

It combines restaurant management with OpenAI to help owners make better decisions.

The system currently allows users to:

  • Manage inventory in real time.
  • Track ingredients and stock value.
  • Detect products below the minimum stock.
  • Manage recipes and food costs.
  • Register sales.
  • Track employee information.
  • Monitor restaurant finances.
  • Calculate inventory investment.
  • Ask an AI assistant questions using live restaurant data.

Instead of searching through menus or spreadsheets, the owner can simply ask:

"Which ingredients are below the minimum?"

or

"What should I buy first and why?"

The AI analyzes the current database and generates an immediate recommendation.

How we built it

Makondo AI Manager was built using:

  • PHP
  • SQLite
  • HTML
  • CSS
  • JavaScript
  • OpenAI Responses API
  • GPT-5 mini

The application stores all restaurant information inside a SQLite database.

When the user asks a question, the backend collects live information including inventory, recipes, sales and purchases.

Those data are sent to the OpenAI Responses API, which generates contextual recommendations specifically for that restaurant.

The interface was designed to remain simple enough for small businesses while providing powerful AI capabilities.

Challenges we ran into

One of the biggest challenges was giving the AI enough business context without overwhelming it.

We had to carefully organize inventory, recipes, costs, purchases and sales into structured data before sending them to the model.

Another challenge was designing an interface that restaurant owners could use without any technical knowledge.

Accomplishments that we're proud of

We are proud that Makondo AI Manager is not a demo with fake data.

It works using real information from an operating restaurant.

The AI answers questions based on the actual inventory, recipes and business data instead of static examples.

This makes every response useful for daily operations.

What we learned

This project taught us that large language models become much more valuable when connected to real operational data.

Instead of replacing existing management software, AI becomes a layer of intelligence capable of helping owners make faster and better decisions.

What's next for Makondo AI Manager

Our roadmap includes:

  • Automatic purchase orders.
  • AI demand forecasting.
  • Staff scheduling optimization.
  • Voice conversations with the AI assistant.
  • Supplier integration.
  • Multi-restaurant support.
  • Mobile applications.
  • Predictive financial analysis.

Our vision is to become an AI operating system for independent restaurants.

Built With

Share this project:

Updates