Google extended Nano Banana, its image generator, with Personal Intelligence — the ability to pull context from a user's connected Google Calendar, Drive, Gmail, and Photos to produce images that reflect the user's actual life rather than generic scenes.
With permission, Nano Banana can now ground a generation in real context:
'Make a birthday card for my dad' becomes a card themed around the photos and references it finds in Photos and past Gmail threads.
'Create a poster for our trip next month' pulls the destination, dates, and travel companions from Calendar and related Gmail confirmations.
'Generate a hero image for the pitch deck' can reference brand assets stored in Drive, including logos and past decks.
Under the hood, Google frames this as a Personal Intelligence layer: the image model reads from a context graph of the user's Google services via scoped, revocable connectors. The user remains in control — each source is opt-in, and Google exposes a dashboard showing what data was used to ground any specific image.
The result is a style of image generation that feels less like prompting a stranger and more like asking a designer who actually knows you. It also escalates the competitive pressure on closed image models that don't have a comparable grounding layer — the gap is no longer only about fidelity, but about how well the model knows the person asking.