In the rapidly evolving landscape of artificial intelligence, Apple finds itself in an unusual position playing catch-up. The tech giant that once pioneered voice assistants with Siri in 2011 now scrambles to reinvent its aging assistant in the era of generative AI. Recent reports reveal the company’s struggles and strategic missteps as it attempts to breathe new life into Siri through a complete LLM-powered overhaul.
The stakes couldn’t be higher. As competitors like Google, Microsoft, and OpenAI push the boundaries of what AI assistants can do, Apple’s once-revolutionary Siri has fallen behind. But transforming the decade-old assistant isn’t proving easy, with internal conflicts, technical challenges, and shifting priorities hampering progress.
The Late Start That Cost Apple Its Edge
Apple’s journey into generative AI began with a significant disadvantage timing. According to Bloomberg’s Mark Gurman, Apple Intelligence “wasn’t even an idea” before ChatGPT launched in late 2022. This late start put Apple years behind competitors who had already invested heavily in large language models and generative AI capabilities.
“In the world of AI, you really don’t know what the product is until you’ve done the investment,” one longtime Apple executive told Gurman. “That’s not how Apple is wired. Apple sits down to build a product knowing what the endgame is.”
This fundamental mismatch between Apple’s product development philosophy and the experimental nature of AI research created a perfect storm of delays and indecision. While other tech giants poured billions into AI infrastructure and talent, Apple hesitated.
Craig Federighi, Apple’s software chief, was reportedly “reluctant to make large investments in AI,” viewing it as a distraction rather than a “core capability.” This reluctance meant fewer GPUs purchased and less aggressive hiring in AI compared to rivals decisions that would later haunt the company as it tried to accelerate its AI efforts.
Leadership Struggles and Cultural Misalignment
The challenges weren’t limited to timing. John Giannandrea, Apple’s AI chief who joined from Google in 2018, faced an uphill battle integrating into Apple’s tight-knit executive team. As a rare outside hire at the executive level, Giannandrea reportedly didn’t “fit in” with Apple’s inner circle and lacked the “forceful” personality needed to secure substantial funding for AI initiatives.
Giannandrea’s perspective on AI also clashed with market realities. He reportedly told employees that customers commonly want to disable tools like ChatGPT. a view that seemed increasingly out of touch as generative AI exploded in popularity.
The cultural disconnect extended beyond personalities. Apple’s methodical, perfection-oriented approach to product development proved ill-suited for the fast-moving, experimental world of AI. While competitors openly tested and rapidly iterated, Apple’s secretive approach proved inadequate for developing AI capabilities..
This spring, Giannandrea was reportedly taken off product development, Siri, and robotics projects. According to Gurman, Apple executives have discussed putting him “on a path to retirement,” though concerns about losing the research and engineering talent he brought with him have complicated matters.
The Technical Nightmare: Merging Old and New
Perhaps the most significant challenge in revamping Siri has been technical. Apple’s initial strategy bolting generative AI features onto the existing Siri architecture proved disastrous.
“It’s whack-a-mole. You fix one issue, and three more crop up,” an Apple employee told Gurman, describing the integration challenges.
The fundamental problem lies in Siri’s architecture. Built long before the era of large language models, Siri relies on a rigid, rule-based system that struggles to accommodate the more flexible, context-aware capabilities of modern AI. Attempting to merge these disparate technologies created a technical quagmire that delayed promised features and frustrated users.
In early 2025, Federighi reportedly tested iOS 18.4 and discovered that many highlighted Siri features such as retrieving a driver’s license number via voice weren’t functioning properly. Internal demos had shown pre-recorded videos instead of working technology, leading to indefinite postponement of the release.
The Marketing Misstep: Promising Too Much, Too Soon
Apple’s marketing department didn’t help matters. The company heavily promoted Apple Intelligence features, including an improved Siri, before they were ready for prime time. This created a gap between expectations and reality that has damaged consumer trust and even led to class-action lawsuits alleging false advertising.
The premature announcements also put enormous pressure on engineering teams to deliver capabilities that weren’t technically feasible within the promised timeframes. As features missed their launch dates one after another, the company’s reputation for reliability suffered.
- AI writing tools and summarizations arrived a month and a half after iPhone 16’s launch
- Genmoji launched in December 2023
- Notification upgrades came only in March 2024
- Siri’s much-awaited upgrade has been delayed repeatedly
The New Strategy: LLM Siri and a Fresh Approach
Facing these challenges, Apple has pivoted to a more fundamental solution: rebuilding Siri from the ground up rather than trying to patch the existing system. This initiative, internally dubbed “LLM Siri,” represents a complete architectural overhaul that will “entirely build on an LLM-based engine.”
According to Gurman, Apple has its AI team in Zurich working on this new architecture, with the goal of making the assistant “more believably conversational and better at synthesizing information.” This approach acknowledges that incremental improvements to the existing Siri won’t be sufficient to compete in today’s AI landscape.
The company is also exploring innovative approaches to data collection and training. One strategy involves leveraging iPhones and differential privacy to improve Apple’s synthesized data comparing fake training data with language from iPhone users’ emails, but doing so on-device and sending only the synthesized data back to Apple for AI training.
Another potential capability being discussed is allowing the LLM version of Siri to access the web to “grab and synthesize data from multiple sources,” essentially turning Siri into an AI web search tool similar to Perplexity.
What to Expect at WWDC 2025 and Beyond
Despite these ambitious plans, significant Siri upgrades are “unlikely to be discussed much” at this year’s WWDC in June, according to Gurman. The promised features from last year’s conference remain “months away from shipping,” suggesting Apple is taking a more cautious approach to announcements.
Instead of major Siri announcements, Apple plans to focus on iOS 19 and gradually expand Apple Intelligence capabilities to additional apps. Planned updates include:
- An AI-powered battery optimization tool
- Project Mulberry—a virtual health coach
- A new setting for European users to replace Siri with third-party voice assistants (distinct from Siri’s ChatGPT integration)
In a telling move, Apple is reportedly preparing to separate the Apple Intelligence brand from Siri in its marketing a tacit admission that Siri’s poor reputation could hinder the company’s broader AI messaging.
The Road Ahead: Can Apple Catch Up?
The challenges facing Apple’s AI efforts extend beyond Siri. The company has fewer AI-focused staff and less investment in GPUs compared to rivals. Whether Apple acted too late with too little investment will soon become clear. Apple failed to acquire companies like Mobileye Global Inc., and ultimately shelved many ambitious AI plans, including those for autonomous driving
Some insiders have likened the company’s AI struggles to a “sinking ship.” Apple’s traditional approach meticulously building polished products hasn’t worked well in the fast-moving world of AI. Longtime executive Eddy Cue reportedly warned that Apple’s dominance is at risk, expressing concern that the iPhone could be irrelevant within ten years if the company doesn’t adapt.
Yet Apple still holds significant advantages: vast resources, a loyal customer base, and unparalleled hardware-software integration. If the company can successfully execute its LLM Siri vision and align its development culture with the realities of AI innovation, it may yet reclaim its position as a leader in voice assistants.
The coming months will be crucial. As Apple works to deliver on delayed promises and build a new foundation for Siri, the tech world watches closely.
The Unrealized Vision of Siri
It’s worth remembering that Siri was once a revolutionary product. Introduced just before Steve Jobs’ passing, it represented a bold vision for the future of human-computer interaction. Jobs envisioned Siri as a key user interface that would transform how we interact with technology.
But that vision never fully materialized. Post-launch development focused on basic tasks rather than the seamless “do engine” originally envisioned. As Apple emphasized other machine learning projects like Maps, Photos, and biometric features, Siri’s evolution stagnated.
Now, with LLM Siri, Apple has a chance to fulfill that original promise creating an assistant that truly understands and anticipates user needs. The question is whether the company can overcome its cultural and technical hurdles to deliver on that potential.
As one Apple executive told Gurman, “In the world of AI, you really don’t know what the product is until you’ve done the investment.” For Apple, that investment in technology, talent, and a new approach to development is finally underway. Whether it’s too little, too late remains to be seen.
