Using the latest AI transcription models, we can (finally) use speech as a working and useful interaction pattern. The new AI-powered models understand the spoken word - even in noisy environments - much better than the current generation of Google Assistants and Siris.