The on-stage demo
In the demonstration led by Mike Rockwell during the keynote, Siri AI received a request based on an image seen in an Instagram post: the user asked for directions to the pictured location, and Siri processed both the visual content and geographic context to return a navigable response. According to coverage from CNBC and Engadget, the ability to 'ask Siri questions about what the iPhone camera is seeing' is an integral part of the new Visual Intelligence layer distributed in iOS 27.
What concretely changes
Until iOS 26, Visual Intelligence was a peripheral feature, accessible via Camera Control on iPhone 16 Pro but separate from the natural Siri interaction flow. In iOS 27, the integration is direct: Siri AI can query in real time what the device's visual system is analyzing, whether through the camera or via screenshots, as already documented for macOS Golden Gate. The practical result is an assistant that responds to the user's perceptual context, not just text or voice commands.
The Google comparison
Engadget's coverage notes that these features are reminiscent of visual search capabilities Google has offered for years via Google Lens. Apple arrives late, but with a structural advantage: deep integration with the native app ecosystem and local data handling via Private Cloud Compute. The differential value is not in the novelty of the feature itself, but in the ability to execute it with a level of privacy by design that Google, by definition, cannot match.