Rafael Schaad

The talk reviews emerging AI interfaces that go beyond chat, focusing on how software can help people control AI driven workflows, agents, and generated content. Through demos of voice, workflow, research, design, email, and video products, the speakers discuss feedback, latency, trust, and ways to keep people involved. They highlight interface patterns such as visual workflow canvases, source links, adaptive controls, keyboard shortcuts, and low fidelity previews during long generation tasks. The central lesson is that AI interfaces need to make actions and progress understandable while giving people useful ways to guide and verify the results.