Siri AI Is Here - What's New, What It Can Do, and Who Can Try It (2026)
Apple released the next generation of Apple Intelligence on September 14, 2026. The headline is a rebuilt assistant called Siri AI. It ships inside Apple's 2027 software releases: iOS 27, iPadOS 27, macOS 27, watchOS 27, and visionOS 27.
The rollout is narrower than the announcement. The beta is English-only at launch. It is not available in the EU or in China. This piece covers what Apple actually described, which devices qualify, and where a phone-resident assistant sits next to the tools you use for work.
What Apple shipped on September 14
Apple's newsroom post calls Siri AI "a completely reimagined version of Siri that is more personal and powerful." Three capabilities carry most of the weight.
The first is personal context understanding. The assistant can search your own messages, emails, and photos to answer a question about something you already have. Apple gives an example. A user asks what a family member suggested for an upcoming visit. Siri finds the message where they mentioned baking a family recipe together, then pulls up the instructions.
The second is systemwide app actions. Apple says these now cover tasks "such as drafting an email from scratch." Siri AI also works with third-party apps already on your device. Apple names WhatsApp and Audible.
The third is onscreen awareness. The assistant can answer questions or take actions about whatever is on your display. Apple's example is a sports team's website. You can ask Siri to find the upcoming home games and add them directly to your calendar.
Any response can be extended into a longer conversation with follow-up questions. How you invoke the assistant also changed. Alongside saying "Hey Siri," iPhone users can use the side button. On Apple Vision Pro, Apple describes a 3D visualization that users place anywhere in their space. You invoke it by looking at it and starting to speak.
There is also a dedicated Siri app. It stores conversation history and syncs it through iCloud. A conversation started on iPhone can continue on Mac, iPad, Apple Watch, or Apple Vision Pro. Apple frames this as bringing rich conversations together in one place across a user's products.
Writing Tools, dictation, and Visual Intelligence
Writing Tools are now integrated across the system. You can describe what you need and Siri will generate a draft from scratch. Describe a change and it updates the text. In Mail and Messages, Apple says Siri can reflect how you usually write to a specific recipient, down to punctuation and tone.
Proofreading runs automatically as you type. Apple says that extends across the system, including inside most third-party apps.
Dictation improved alongside it. Apple credits its most advanced on-device model, AFM Core Advanced. Dictation now handles capitalization, punctuation, and formatting as you speak. Voices are more expressive, and you can adjust both expressiveness and pace.
Visual Intelligence extends to iPad, Mac, and Apple Vision Pro. On iPhone it lives in the Camera app as a new Siri mode. On iPad it hooks into the screenshot flow. On Mac a keyboard shortcut lets you select something on screen and type a question about it.
What runs on device and what runs in the cloud
Apple describes a split architecture. Some work happens on device. The rest goes to Private Cloud Compute. Apple says that tier gives users "access to frontier-level intelligence while extending the privacy and security of iPhone into the cloud."
Apple states that personal data is not stored when Private Cloud Compute handles a request. It is not made accessible to Apple or anyone else. Apple adds that outside experts can continue to verify this privacy promise at any time.
One detail is easy to miss. Apple credits "the next generation of Apple Foundation Models, custom-built in collaboration with Google and its Gemini models." Apple says those models run on device and on servers using Private Cloud Compute. The assistant is Apple's. Part of the model work behind it is a partnership.
Siri AI also leans on a system orchestrator. That reaches the Spotlight index and something Apple calls App Toolbox. Apple says both work entirely on device.
Who can try Siri AI right now
This is where the announcement gets specific, and where most practical answers live.
| Requirement | What Apple stated |
|---|---|
| Software | iOS 27, iPadOS 27, macOS 27, watchOS 27, or visionOS 27 |
| iPhone | iPhone 16 or later, plus iPhone 15 Pro and iPhone 15 Pro Max |
| iPad | iPad mini (A17 Pro), and iPad models with M1 or later |
| Mac | MacBook Neo (A18 Pro), and Mac models with M1 or later |
| Watch | Series 9 or later, Ultra 2 or later, SE 3 with a nearby enabled iPhone |
| Other | Apple Vision Pro |
| Language | English beta now; French, Japanese, Korean, Portuguese, Spanish next month |
| Not available | EU on iOS, iPadOS, and watchOS; China |
| Usage | Daily caps on server-side features |
A few of those rows deserve expanding.
On language, Apple Intelligence as a whole supports sixteen. Apple lists Chinese in both simplified and traditional forms, plus Danish, Dutch, English, French, German, Italian, Japanese, Korean, Norwegian, Portuguese, Spanish, Swedish, Turkish, and Vietnamese. That wider list is not the same as the Siri AI beta list. The assistant itself starts in English. Apple also notes that some features may not reach every region or language, and that some devices are not sold everywhere.
On regions, Apple states plainly that Siri AI "will not be available initially in the EU on iOS, iPadOS, and watchOS." Features relying on it are also unavailable there. Apple says it is working to find a path forward that preserves user privacy and security. For China, Apple cites regulatory requirements it is still working through.
On limits, features that depend on server-side models carry daily usage caps. Apple names the new assistant among them. Apple also writes that "expanded access to such features will be available for a fee in the future." Today's caps are the free tier of something that will eventually have a paid one.
Two more features, Siri Recap and Live Rewind, arrive in beta later this year rather than today.
What changed outside the assistant
The release touched a long list of built-in apps. Photos gained Spatial Reframing, which adjusts camera perspective after the shot. Image Playground now produces photorealistic styles. Safari can group tabs into topics automatically. A feature called Notify Me watches a page for changes such as restocks or price drops. Mail's Smart Reply draws on your writing style. Shortcuts now lets you build an automation by describing it.
The Home app can now treat related notifications as a single activity. It also generates descriptions across a sequence of video clips, and searches for clips across multiple cameras.
The Health app was redesigned around the same models. A new Insights tab carries an Apple Intelligence summary of a user's most timely information. A new Longevity tab analyzes longitudinal health data, including a metric Apple calls Health Age.
On Apple Watch, a new category called Audio Intelligence covers Sound Recognition, Live Rewind, Siri Recap, and background Shazam. Apple is explicit about the privacy design. These features "do not create or store audio recordings." They do not identify or attribute speakers. Siri Recap produces "a brief, high-level summary of conversations — not a transcript." Live Rewind snippets and Siri Recap summaries are end-to-end encrypted, and Apple says they are not accessible even by Apple.
Where a device assistant fits in a workday
Read Apple's examples closely and a pattern shows up. The assistant works across messages, emails, photos, calendar entries, and whatever is on your screen. Its raw material is the information already flowing through your devices. Its output is usually an action: a draft, a calendar entry, a reply in the right tone.
That covers a real and large part of a workday. A different part starts when the question changes shape. "What did someone tell me" is one kind of question. "What does this pile of files actually say" is another.
Think about a quarter of raw exports, a set of survey responses, or three spreadsheets that disagree with each other. That work begins with files you bring to it. It ends in a chart, a written finding, or a deck someone else has to read.
That second shape is what Powerdrill Bloom is built around. You upload the files and describe in natural language what you want to know. The agent works through the analysis, the visualization, and the output. It runs the same way whether the source is an Excel export, a CSV, a PDF, or a folder of documents. It returns slides, docs, and sheets rather than a spoken answer.
The two are not competing for the same moment. One is the assistant you talk to while holding your phone. The other is where a file-shaped question goes. If your week involves both, it helps to know which one you are reaching for.
For context on how other vendors framed launches this month, our write-up of DeepSeek V4.1-Flash covers the same questions for an open-weights model release. If you are weighing general assistants against tools built for reporting work, our roundup of AI business intelligence tools draws that line. It covers where each kind of tool stops.
Ready to see what a file-first agent does with your data? Start with Powerdrill Bloom.
Frequently asked questions
What is Siri AI?
Siri AI is Apple's rebuilt assistant, released September 14, 2026. It is powered by the next generation of Apple Intelligence. Apple describes it as "a completely reimagined version of Siri" with personal context understanding, systemwide app actions, and onscreen awareness.
How is the new Siri different from the old one?
Apple says Siri was "rebuilt from the ground up." In practice it can now search your own messages, emails, and photos for context. It can act inside apps on your behalf. It can answer questions about what is on your screen. Conversations continue across follow-up questions in a dedicated app.
Why can't I use Siri AI in the EU?
Apple's announcement states that Siri AI "will not be available initially in the EU on iOS, iPadOS, and watchOS." Apple says it is working to find a path forward that preserves its users' privacy and security. For China, Apple gives regulatory requirements as the reason.
Which devices and languages are supported?
You need iOS 27, iPadOS 27, macOS 27, watchOS 27, or visionOS 27 on an Apple Intelligence-enabled device. The table above lists the qualifying hardware. The assistant launches in English beta, with French, Japanese, Korean, Portuguese, and Spanish scheduled for next month.
Is Siri AI free?
There is no separate price today. Apple notes that server-side features have daily usage limits, and that "expanded access to such features will be available for a fee in the future."
Conclusion
Apple shipped a genuinely rebuilt assistant, then drew a tight border around who gets it first. English, beta, no EU, no China, daily caps on anything touching a server. If you are inside that border and on a qualifying device, start with the personal context and onscreen features. Those are the ones the old Siri could not reach.
The interesting part is the split it makes visible. Assistants living on your device keep getting better at the information you already have. Work that starts with files you bring still belongs somewhere else. Knowing which question you are asking is most of the decision.