AI agents are changing how users interact with software, yet most demonstrations happen on desktop or in the cloud. How do we bring these intelligent agents to mobile, where resources are constrained and user expectations are high?

This talk explores the practical reality of building AI agents for mobile apps. Rather than focusing on running LLMs locally, which remains impractical for most applications, I'll show you how to build agents that combine on-device intelligence with lightweight cloud orchestration to give you the best of both worlds.

We'll cover:

1. Architecting mobile AI agents
How to structure your app so AI features feel native rather than bolted on.

2. Tool calling on mobile
How to give AI agents the ability to perform actions inside your app, from filling forms and navigating screens to fetching data.

3. Intelligent offline capabilities
How to use on-device models for tasks that don't need the cloud, such as text classification and entity extraction.

4. Streaming interactions
How to build chat like experiences that feel responsive even on mobile networks.

I'll walk through a real world example: a smart expense tracking agent that can extract receipt data, categorize spending, and answer questions about your finances, all while respecting mobile constraints around battery, data, and privacy.

Attendees will leave with practical patterns they can apply immediately to add AI agents to their Android or iOS apps.