Apple Reboots Siri with AI Overhaul Amid Internal Struggles

Apple is making a bold push to revamp its artificial intelligence efforts, with a renewed focus on transforming Siri into a cutting-edge, conversational assistant powered by large language models (LLMs). The move follows a rocky debut of its Apple Intelligence features, which faced delays, internal resistance, and strategic missteps.

In a detailed report by Bloomberg’s Mark Gurman, the inner workings of Apple’s AI development are laid bare — revealing why the tech giant fell behind in the AI race and how it plans to catch up. At the heart of the recovery plan lies a complete rebuild of Siri, internally dubbed “LLM Siri,” which aims to move past the limitations of its legacy code and become a more natural, responsive digital assistant.

According to Gurman, part of the problem stemmed from hesitation at the top. Craig Federighi, Apple’s senior vice president of software engineering, was reportedly “reluctant to make large investments in AI.” The company has historically shied away from funding projects without clear, defined outcomes. However, when it comes to AI, one unnamed executive told Gurman, “...you really don’t know what the product is until you’ve done the investment.”

This caution delayed Apple’s acquisition of critical GPU resources needed to develop powerful AI systems. By the time Apple recognized the urgency, competitors like OpenAI and Google had already secured a lead.

Another issue was timing. Apple Intelligence, according to insiders, “wasn’t even an idea” before OpenAI’s ChatGPT launched in late 2022. This late start left Apple scrambling to catch up — and with limited internal alignment.

John Giannandrea, Apple’s AI chief, reportedly didn’t believe users were interested in AI chatbots. “He told employees that customers commonly want to be able to disable tools like ChatGPT,” Gurman noted. His reserved approach clashed with the growing market demand for interactive, generative AI tools.

Attempting a shortcut, Apple tried layering generative AI onto the existing Siri framework. But this hybrid solution quickly proved unstable. “It’s whack-a-mole. You fix one issue, and three more crop up,” an Apple employee told Gurman.

Giannandrea’s struggles extended beyond strategy. Having joined Apple in 2018 from Google, he remained somewhat on the fringes of Apple’s typically insular leadership. Described as less “forceful” than his peers, he reportedly failed to secure the funding and urgency needed to accelerate AI development. Employees also claimed he didn’t push his team hard enough and underestimated rivals like OpenAI and Google.

Meanwhile, Apple’s marketing got ahead of reality. The company heavily promoted features like a context-aware Siri and Apple Intelligence’s ability to analyze data across apps — capabilities that were not yet fully developed. This forced delays and scaled-back expectations.

Now, Apple is resetting its course. The revamped version of Siri, being built in Zurich, is expected to run entirely on an LLM-based engine. This promises a more fluid, conversational interface, capable of understanding context and synthesizing information effectively. Gurman previously reported on this effort last November, and momentum appears to be growing.

In addition to technical upgrades, Apple is also exploring ways to improve AI training data while preserving privacy. One method involves comparing synthetic training data against user email language locally on iPhones, only sending anonymized, synthesized data back to Apple.

Furthermore, Apple is considering turning LLM Siri into a web-savvy assistant — one that can “grab and synthesize data from multiple sources,” much like AI search tools such as Perplexity. Apple has reportedly explored partnerships to integrate such capabilities into Safari.

Notably, Giannandrea is stepping away from product development, Siri, and robotics initiatives. Gurman reports that Apple executives have even discussed a potential retirement path for him, although there are concerns that key researchers and engineers might leave with him. Giannandrea, for his part, appears to be staying put — reportedly “relieved Siri is now someone else’s problem.”

Apple’s AI efforts may have stumbled out of the gate, but with a complete reboot of Siri and a clearer vision, the company is working hard to reclaim its edge in the fast-evolving world of artificial intelligence.