Apple’s rebuilt Siri ships as an opt-in beta with limits

If you ship an iOS or macOS app, the decision is whether to expose functionality through App Intents now, while the assistant is a capped beta. Exposing intents is cheap and reversible; making the assistant the only route to a feature is neither. Assume absence as the default case: EU users on iPhone, iPad and Watch do not have it, mainland China does not have it, non-English users do not have it yet, and anything below the supported device floor never will. Build the voice path as an accelerator over an existing UI, not a replacement for it. Because server-backed features carry undisclosed daily limits, treat a refusal or a fallback as an expected state to handle, not an error to log and ignore.

The eligibility question comes before the enthusiasm. A fleet policy that assumes every employee gets the same assistant is wrong on day one: Apple’s own launch terms exclude EU iPhone, iPad and Watch users and mainland China, and gate the rest on device generation and language. If you operate in regulated work, the data-handling claim is Apple’s statement about Private Cloud Compute and its use of Google Gemini models, not a certification you can hand to an auditor — ask for it in writing and check what your MDM controls can actually disable. On cost, Apple has said a paid tier is coming but published no price, so any business case that assumes a future rate is assuming a number nobody has published.