ZDNet editor puts the new Siri with Apple Intelligence through 10 tasks on a Mac

The new Siri with Apple Intelligence is making its debut on the Mac, and the first independent tests paint a mixed picture. The assistant now understands conversational context, can hand complex requests off to ChatGPT, and works more deeply inside apps. It isn't all smooth, though: Siri handles some tasks confidently, while others come back with errors or delays. It's a promising start, but Apple still has serious work ahead.

AI-processed from ZDNet AI; edited by Hamidun News
ZDNet editor puts the new Siri with Apple Intelligence through 10 tasks on a Mac
Source: ZDNet AI. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceZDNet AI ↗
Editor-in-Chief
Key metric #1Contextual search — the assistant finds the right email or reminder without misses, drawing on the actual contents of mail and notes
Key metric #2System commands — switching Focus modes, managing files, changing app settings
Key metric #3Writing Tools — proofreading, rewriting, and summarizing selected text work reliably

Apple Intelligence, the new AI layer on top of macOS Sequoia, has finally reached a broad audience of Mac users worldwide, and it brings a radically reworked Siri with it. A ZDNet editor went beyond first impressions and put the assistant through 10 hands-on tests built around real everyday tasks. The verdict is mixed, and revealing.

What fundamentally changed

The biggest difference in the new Siri is deep access to the user's personal context. The assistant now reads emails in Mail, messages in Messages, events in Calendar, and notes, and it generates answers directly on the device without sending data to the cloud. Requests like "When did I last talk to this person?" or "Remind me about the meeting today at three" no longer leave Siri in confused silence.

Alongside this, Apple has integrated OpenAI's ChatGPT for requests that go beyond personal data. Users decide each time whether to allow a request to be passed to the external system.

The interface has changed too: the new Siri glows around the edges of the screen instead of showing the old bubble, accepts typed input alongside voice, and can keep track of context across several exchanges in a row.

Where the assistant delivered

The test showed that the updated Siri performs confidently wherever it needs deep integration with the system and the user's personal data.

  • Contextual search — the assistant finds the right email or reminder without misses, drawing on the actual contents of mail and notes
  • System commands — switching Focus modes, managing files, changing app settings
  • Writing Tools — proofreading, rewriting, and summarizing selected text work reliably
  • Notification summaries — long conversations collapse into a prioritized one-line summary
  • Voice tasks — reminders and events get created quickly and accurately, even from vague phrasing

Where there's still work to do

Not everything went off without a hitch. Some requests take longer than you'd expect from an on-device assistant: a pause of several seconds gets irritating in everyday use. Siri occasionally misreads complex compound commands, especially when a task touches two or three apps at once.

Several Apple Intelligence features are still marked "coming soon": full integration with third-party apps and an agent mode that carries out multi-step tasks without constant confirmation remain in development. Users can see the promise, but not always the actual capability.

"This is a promising start, but

Apple still has work to do," the tester concludes.

What it means

Apple Intelligence isn't a one-time release but a living product that Apple is rolling out in waves of updates over several months. Upcoming versions of macOS Sequoia could significantly change the picture, especially when it comes to autonomous agents and an expanded API for third-party developers.

The AI assistant race between Apple, Google, and Microsoft is becoming more visible every month. For users who have long waited for a genuinely useful voice assistant inside the Apple ecosystem, the new Siri is a meaningful step forward. For those who hoped to get a ready-made rival to GPT agents on the Mac right away, the wait continues for now, but Apple has picked the right direction.

Story

Key Questions Answered

What fundamentally changed?

The biggest difference in the new Siri is deep access to the user's personal context. The assistant now reads emails in Mail, messages in Messages, events in Calendar, and notes, and it generates answers directly on the device without sending data to the cloud. Requests like "When did I last talk to this person?" or "Remind me about the meeting today at three" no longer leave Siri in confused silence. Alongside this, Apple has integrated OpenAI's ChatGPT for requests that go beyond personal data. Users decide each time whether to allow a request to be passed to the external

Where the assistant delivered?

The test showed that the updated Siri performs confidently wherever it needs deep integration with the system and the user's personal data. Contextual search — the assistant finds the right email or reminder without misses, drawing on the actual contents of mail and notes System commands — switching Focus modes, managing files, changing app settings Writing Tools — proofreading, rewriting, and summarizing selected text work reliably Notification summaries — long conversations collapse into a prioritized one-line summary Voice tasks — reminders and events get created quickly

Where there's still work to do?

Not everything went off without a hitch. Some requests take longer than you'd expect from an on-device assistant: a pause of several seconds gets irritating in everyday use. Siri occasionally misreads complex compound commands, especially when a task touches two or three apps at once. Several Apple Intelligence features are still marked "coming soon": full integration with third-party apps and an agent mode that carries out multi-step tasks without constant confirmation remain in development. Users can see the promise, but not always the actual capability. "This is a promis

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…