Testing Apple's New Siri AI Assistant: Does It Actually Work?

After years of being the punchline of every virtual assistant joke, Siri has finally gotten a massive overhaul. Apple's new AI-powered Siri — built on the foundation of Apple Intelligence and rolled out progressively since late 2025 — promises contextual awareness, on-screen understanding, conversational memory, and deep app integration. But does it actually deliver? We spent three weeks putting the new Siri through rigorous, real-world testing across an iPhone 16 Pro, iPad Pro M4, and MacBook Air M3 to find out exactly where it shines and where it still stumbles.

What's Actually New About Siri in 2026?

Before diving into results, it's worth understanding what Apple changed. This isn't the old command-and-response Siri with a fresh coat of paint. Apple fundamentally rebuilt Siri's architecture using large language models integrated through Apple Intelligence, which the company first introduced at WWDC 2024 and has been expanding ever since.

Here are the major upgrades now live as of mid-2026:

• Conversational memory: Siri can remember context across multiple exchanges in a single session and, in some cases, across sessions.
• On-screen awareness: Siri can see and understand what's currently displayed on your screen and take actions based on it.
• Deep app actions: Rather than just opening apps, Siri can now perform complex multi-step tasks inside third-party apps.
• Personal context: Siri draws from your emails, messages, calendar, photos, and files to give personalized answers — all processed on-device for privacy.
• Natural language understanding: You no longer need to use robotic phrasing. Siri handles rambling, corrections, and follow-up questions far better.

According to a Bloomberg Intelligence report from April 2026, early adoption data suggests that Siri usage has increased approximately 68% since the AI overhaul began, with the average session length nearly tripling compared to pre-update baselines.

Our Testing Methodology

We tested the new Siri across five categories that matter most to everyday users:

1. General knowledge and conversation
2. Personal context and scheduling
3. On-screen actions
4. Third-party app integration
5. Multi-step complex tasks

Each category included at least ten distinct prompts, ranging from simple to highly complex. We scored Siri on accuracy, speed, and whether it completed the task without requiring manual intervention.

The Results: Where Siri Actually Excels

Personal Context Is Genuinely Impressive

This is where the new Siri truly shines. We asked things like:

• "When is my dentist appointment next week?"
• "What was the restaurant Sarah recommended in her text last Thursday?"
• "Show me the photos I took at the beach last month."

Siri nailed these consistently. It pulled calendar entries, scanned through Messages, and surfaced the right photos with remarkable accuracy. The fact that all of this happens on-device — Apple's private cloud compute handles the heavy lifting without sending your personal data to external servers — makes it feel both powerful and trustworthy.

Conversational Flow Feels Natural

We tested back-and-forth conversations like:

• "What's the weather in Tokyo?"
• "What about Osaka?"
• "Will I need an umbrella on Friday?"

Siri maintained context beautifully across all three prompts without needing us to repeat "weather" or "Japan." This sounds basic, but it's something old Siri would have fumbled after the second question. In 2026, it feels like talking to someone who's actually listening.

On-Screen Awareness Is a Game-Changer

We opened a webpage with a recipe and said, "Add these ingredients to my grocery list in Reminders." Siri read the page, identified the ingredient list, and created the reminder entries. We tried the same with a restaurant page — "Save this address to my contacts" — and it worked flawlessly.

This feature alone justifies the upgrade for many users.

Where Siri Still Falls Short

Third-Party App Integration Is Hit or Miss

Apple has opened its App Intents framework for developers, but adoption remains inconsistent. Siri worked well with major apps like Spotify, WhatsApp, and Uber. For example, "Play my Discover Weekly on Spotify" and "Send a WhatsApp message to Mom saying I'll be late" both worked without issue.

However, when we tried more niche apps — a specific workout tracker, a note-taking app outside Apple's ecosystem, a project management tool — Siri either defaulted to opening the app without completing the task or apologized and offered a web search. Apple is clearly dependent on developers adopting the new APIs, and not everyone has caught up yet.

Complex Multi-Step Tasks Are Unreliable

We pushed Siri with requests like:

• "Find the cheapest flight to Denver next weekend, then add the departure time to my calendar and text the itinerary to Jake."
• "Summarize the last three emails from my boss and draft a reply to the most recent one."

Siri handled parts of these tasks but rarely completed the entire chain without stumbling. The email summarization worked well, but drafting a contextually appropriate reply was inconsistent — sometimes impressive, sometimes generic to the point of being useless. The flight request fell apart entirely, as Siri doesn't yet have deep booking integration.

Occasional Hallucinations and Errors

While significantly improved, Siri still occasionally provides confidently wrong information — a hallmark issue of LLM-based systems. When asked about recent news events, it twice gave us slightly inaccurate details. For factual queries where precision matters, we'd still recommend verifying independently.

How Does It Compare to the Competition?

The AI assistant space in 2026 is fiercely competitive. Google's Gemini assistant leads in web-based knowledge and real-time information. ChatGPT's integration across devices offers the deepest conversational AI experience. Amazon's Alexa has reinvented itself with its own LLM upgrade.

Here's how they stack up in our testing:

| Feature | Siri (2026) | Google Gemini | ChatGPT | Alexa | |---|---|---|---|---| | Personal context | ★★★★★ | ★★★★ | ★★★ | ★★★ | | Privacy | ★★★★★ | ★★★ | ★★★ | ★★★ | | General knowledge | ★★★★ | ★★★★★ | ★★★★★ | ★★★ | | Third-party apps | ★★★ | ★★★★ | ★★★ | ★★★★ | | Complex tasks | ★★★ | ★★★★ | ★★★★ | ★★★ |

Siri's advantage is clear: it lives inside the Apple ecosystem and understands your data better than anyone while keeping it private. Its weakness remains breadth — it simply can't do as many things outside Apple's walled garden.

Should You Rely on the New Siri?

Here's our honest take after three weeks of daily testing:

• Use it for: personal scheduling, quick lookups, on-screen actions, messaging, reminders, Apple ecosystem tasks, and casual conversation.
• Don't rely on it for: complex research, multi-step workflows involving third-party services, or anything requiring factual precision without verification.
• Best on: iPhone 16 series and newer iPads/Macs with Apple Silicon, where on-device processing makes responses noticeably faster.

The Bottom Line

Apple's new Siri is, without exaggeration, the biggest leap the assistant has made since its launch in 2011. It's no longer embarrassing. It's genuinely useful for everyday tasks, and the on-device personal context feature is something no competitor matches at this level of privacy. But it's not the all-knowing, do-everything assistant Apple's marketing might suggest. Third-party integration gaps and inconsistencies with complex requests keep it from being a true replacement for manual effort.

If you're deep in the Apple ecosystem, the new Siri is worth embracing — just don't throw away your keyboard yet. The foundation is strong. Now it's up to Apple (and its developers) to finish building on it.