Report outlines how Siri is holding up Apple’s entire product launch strategy

Macworld

Apple’s Siri voice assistant is something of a punching bag in the tech industry, with numerous articles pointing out its inaccuracy and lack of cutting-edge features. (Even Apple itself has internally acknowledged that delays to the rollout of new features are “ugly and embarrassing.”) But a new report reveals that these shortcomings are more than just an embarrassment: they are holding up the company’s entire product launch strategy.

The issue, as Bloomberg reporter Mark Gurman writes in the latest edition of his Power On newsletter, is that Apple needs far more from Siri than it’s currently delivering. And it can’t push forward with these plans while the software remains so flawed.

“The real game changer,” Gurman explains, “is an upgraded version of App Intents that could finally make Siri the true hands-free controller of your iPhone.” And that has to wait for everything else.

App Intents is an umbrella term for the framework that allows app developers to integrate their software into Apple’s operating systems and have it interact with Spotlight search, Shortcuts, widgets, the iPhone’s Action button, and Siri voice commands. At the moment, Siri’s ability to operate within apps, particularly third-party apps, is fundamentally limited. But in the near future, Apple will enable Siri to find, edit, and share a photo; scroll through a shopping app, find a product, and add it to your cart; log into a service via voice alone, and leave comments on social media, among many, many other things.

This is the sort of agentic power that is often discussed as the key to transformative AI. If AI can only respond to voice or text with voice or text, there’s a limit to how much time it can actually save you. But if it can understand complex and practical tasks, such as researching and making a purchase, planning and booking a vacation, or creating and sending out party invitations, then responding to the RSVPs, then it becomes the sort of virtual butler sci-fi movies have always promised us.

Siri’s App Intents feature was initially supposed to be part of Apple Intelligence in iOS 18, but development roadblocks forced Apple to delay the release of the new Siri until iOS 26. Apple hasn’t said when it will arrive, but rumors point to iOS 26.4 in early 2026 as the most likely

It’s exciting that this is finally in the cards for Siri, as it will open up a vista of new opportunities. Next-gen App Intents are key to a battery of smart home products which Apple has in the pipeline and which it has delayed several times in order to wait for Siri to catch up, Gurman reports. “Without the new App Intents,” he writes, “those products would potentially be even less compelling than the devices that [Amazon] and [Google] launched half a decade ago. That’s why the Siri delay has rippled through the company’s other product plans.”

Tim Cook can’t simply snap his fingers and have Siri become the agentic voice assistant it needs to be. The problem is the part where I said we need to be able to rely on the AI doing what it’s told to an acceptable standard: a one in 20 failure rate might be OK when playing songs on the HomePod or generating emojis, but it needs to be far more accurate when making purchases or dealing with health matters. So Apple needs to make Siri accurate, and give it the ability to draw on context when responding to commands, and gain agentic power, before it can do everything else it wants to.

One alternative to this roadmap is for the company to ringfence next-gen App Intents to a limited array of apps with which it’s been carefully tested (Gurman mentions “Uber, AllTrails, Threads, Temu, Amazon, YouTube, Facebook, WhatsApp, and even a few games”), and/or exclude sensitive app categories entirely. But agentic AI is hardly agentic AI if you have to memorize the list of scripted tasks it can and can’t do. And Apple will therefore have to choose between an unflattering comparison with the far more ambitious AI released by other companies, and the risk of major Siri mistakes resulting in something far more serious than an ability to say what month it is.