Apple’s big bet on consumer AI hinges on a smarter, more useful Siri—and that upgrade may be hitting turbulence. According to reporting from Mark Gurman, internal testers of the next-generation Siri are raising performance concerns ahead of its planned debut with iOS 26.4, expected in spring 2026. It’s still months away, but the chatter suggests Apple’s most ambitious voice assistant overhaul in years isn’t out of the woods yet.
The company has been rebuilding Siri to feel less like a command line and more like a capable digital aide. Earlier reports indicated engineers were wrestling with reliability and consistency across apps, as well as performance in high-stakes scenarios such as banking. Those challenges appear to be lingering in early dogfooding, even as Apple pushes toward launch.
What the new Siri aims to deliver
– In‑app actions: Execute context-aware tasks inside supported apps using voice. Think “add olive oil to my Costco list,” “send that last photo to Dad,” or “play the latest album from my library” without manually tapping through menus.
– Personal context awareness: Use on-device understanding of your data to surface what you need faster—for example, digging through Messages to find the podcast a friend recommended and opening it directly.
Behind the scenes, Apple’s Answers, Knowledge and Information (AKI) team has also been working on giving Siri the ability to fetch up-to-date information from the web, similar to how modern AI chat systems answer open-ended questions. That capability could significantly expand what Siri can do beyond device-bound tasks.
A high-profile personnel move has added intrigue. Ke Yang, who was recently tapped to lead the AKI group, is reportedly leaving for a new role at Meta. While there’s no official link between the move and Siri’s current state, the timing raises fresh questions about Apple’s AI roadmap and hiring momentum at a critical moment.
Why the architecture shift matters
Apple previously acknowledged that Siri’s first‑generation architecture limited quality and reliability. Software chief Craig Federighi said the company chose to pivot rather than ship something that wouldn’t meet customer expectations, moving to a redesigned “V2” foundation. The new architecture is supposed to unlock more flexible reasoning, deeper app control, and better personalization—all while preserving Apple’s privacy-first approach. But if today’s internal feedback is any indication, even V2 is encountering early growing pains.
What to watch next
– Scope at launch: Apple may choose a phased rollout, limiting certain features to specific apps or regions until performance meets its bar.
– On-device vs. cloud: Expect Apple to keep as much processing on-device as possible, with carefully controlled server-side components for web results.
– Reliability in sensitive tasks: Banking, payments, and account changes will be the ultimate test of trust; Apple is likely to move conservatively here.
– Developer integrations: The usefulness of in‑app actions depends on deep, stable APIs and third‑party support. Watch for guidance to developers in the run-up to release.
The takeaway
Apple is aiming to turn Siri into a genuinely helpful, AI-powered assistant that can act inside apps, understand personal context, and answer questions with current information. Internal testers voicing concerns isn’t unusual for a project of this scale, but it underscores how hard it is to marry generative AI ambition with Apple’s bar for reliability, privacy, and security. If Apple sticks the landing, iOS 26.4 could mark Siri’s most consequential upgrade to date. If not, expect a cautious debut and rapid iteration as the company trains, tunes, and scales its next-gen voice assistant.
