A year ago, a Reddit thread went viral for all the wrong reasons after a user posted their experience with Siri. When they asked it “what month is it?”, Siri responded with “Sorry, I don’t understand.” That post came after a much-hyped Apple’s voice assistant failed to deliver on its most-advertised capabilities.

At WWDC 2024, Apple dazzled audiences with a vision of a new Siri which is contextually aware, deeply personal, capable of reading your screen, acting across apps, and holding a proper conversation. It advertised the iPhone 16 on the strength of these features. People bought those phones expecting that Siri. They didn’t get it.

Craig Federighi, Apple’s SVP for Software Engineering, finally admitted at last year’s developers’ conference that the company tried to merge two fundamentally different Siri architectures and the result was, to put it diplomatically, not stable enough to ship. Class-action lawsuits followed, making the iPhone maker agreed to pay $250 million settlement.

The Siri subreddit became a rolling museum of dysfunction, with posts titled “Siri is a complete disappointment” and “Okay, Siri can’t even set timers anymore” accumulating likes at a depressing rate. CEO Tim Cook, in an earnings call, said Apple was “making good progress.” It was not good enough progress. The product wasn’t ready for iOS 18. It wasn’t ready for iOS 26. The internal target quietly moved to spring 2026.

So when Apple’s executive team walked onto the WWDC 2026 stage on Monday (June 8, 2026) and finally unveiled what the company calls Siri AI, I was all ears. And from what I heard and saw, I was genuinely impressed.

The new Siri AI

Siri AI is a ground-up rebuild on a deep learning architecture, designed for the thing the old Siri never truly had: conversational memory. You can ask a follow-up question without repeating your entire context. You can refine a request mid-thought. The assistant remembers where you’ve been in a conversation, which sounds basic until you remember that the previous Siri treated each utterance like the opening line of a relationship it had never had with you.

The personal context piece is where it gets interesting. Apple Intelligence now draws from your photos, messages, emails, and calendar to give you answers that are grounded in your actual life. Ask about a dinner you have coming up and it knows, because it read your calendar. It can see what’s on your screen. That means if you’re in Maps and ask about somewhere you’re looking at, it understands the context without you needing to spell it out. This on-screen awareness, paired with what Apple is calling an “app toolbox,” means Siri can now perform multi-step actions across apps.

In the WWDC demo, Apple VP Mike Rockwell asked Siri for directions to a landmark seen in an Instagram post. It worked. The visual intelligence capabilities that requires Siri to use the camera or your screen to understand the physical world around you represents something genuinely new in the Apple ecosystem.

And then there’s the privacy architecture, which is the one area where Apple has, throughout all of this turmoil, maintained a meaningful advantage. The combination of on-device processing and Private Cloud Compute means your data isn’t sitting on Apple’s servers, isn’t being used to train future models, and exists only long enough to execute your specific request.

The caveats

While all of this sounds great, the sceptic in me still wants to experience this new Siri AI to know for sure whether Apple has really cracked the AI puzzle.

The company is clearly leaning heavily on Google’s Gemini models to power parts of the experience, a partnership that raises its own questions about Apple’s long-term AI strategy and whether it can ever build the frontier model capabilities in-house that this vision truly demands.

OpenAI’s ChatGPT has had two years to embed itself into people’s daily habits. Google’s Gemini is baked into Android. Even Microsoft’s Copilot has found its footing in enterprise contexts. The people who were going to be early AI assistant adopters have already adopted. Apple is now fighting to convert the mainstream, which is both a harder and a more important battle.

Here is what I actually think: what Apple showcased on Monday are sound. The new conversational engine, the privacy model, and the system-wide integration are genuine upgrades. They also represent a coherent vision of what a personal AI assistant should be, one that is tightly woven into the fabric of your devices rather than arriving as a third-party chatbot you have to consciously switch to.

If it ships as demonstrated, it will be the most personal and, crucially, the most private AI assistant on the market. That is a genuine differentiation in a landscape where the major players are increasingly indistinguishable at the capability level.

The “if” is doing a lot of work in that sentence.

Apple has one more chance to get this right before the narrative hardens into something it can’t recover from. WWDC 2026, Mr. Cook’s final keynote as CEO, was a show of intent. What comes in the fall, and whether the new Siri AI actually works on the phones people actually own, will be the verdict.