Inspiration

Today's AI assistants are incredibly capable but they're trapped inside apps. Every time we want AI to do something, we still have to unlock a device, open the right application, copy information between devices, and manually execute tasks.

I wanted to build an AI that doesn't just answer questions, but actually gets work done.

That idea became Donna: an ambient AI that lives across your phone, laptop, watch, and connected devices. Instead of thinking about which device to use, you simply ask Donna, and it figures out where and how to complete the task.

To make the interaction feel natural, I also built a physical hardware interface. Instead of opening an app, you simply press a button, speak naturally, and Donna begins working immediately.

What it does

Donna is a cross-device AI execution layer.

It can understand natural language, maintain context across devices, and perform actions on the user's behalf instead of only generating text.

Examples include:

Executing tasks on your laptop
Accessing information from your phone
Interacting with your smartwatch
Controlling connected hardware and IoT devices
Coordinating workflows across multiple devices from a single conversation

Rather than being another chatbot, Donna behaves more like an operating system for your digital life.

How I built it

Donna combines software orchestration with physical hardware.

The project consists of:

A hardware interface that provides an always-available way to interact with Donna
A cross-device communication layer connecting multiple devices
An AI orchestration engine that interprets user intent
Device integrations that allow Donna to execute actions where they make the most sense
A unified conversational interface so users don't have to think about individual applications or devices

The system focuses on turning natural language into real-world actions instead of stopping at text responses.

Challenges I ran into

The biggest challenge wasn't building an AI model it was building a system that could coordinate multiple devices reliably.

Some of the main engineering challenges included:

Designing a consistent communication layer between devices
Keeping interactions fast enough to feel natural
Managing permissions across different platforms
Making hardware and software work together seamlessly
Creating an experience that feels like one intelligent assistant instead of several disconnected applications

Another challenge was deciding what to demonstrate. Rather than trying to support dozens of integrations, I focused on building a polished end-to-end experience that clearly shows Donna's potential.

What I learned

This project reinforced that the future of AI isn't just better language models.

The real challenge is giving AI the ability to interact with the world around us.

Building Donna taught me about distributed systems, hardware-software integration, cross-device orchestration, user experience design, and the importance of creating interactions that feel effortless.

I also learned that sometimes the most memorable innovation isn't a bigger model it's a better interface.

What's next for Donna

Donna is just the beginning.

The long-term vision is to make Donna an ambient AI that works across every device people own, understands context, securely accesses the tools users already rely on, and becomes the simplest way to interact with technology.

Instead of opening apps, switching devices, and managing workflows manually, people will simply ask Donna and it will get the work done.