Still no AI-powered, ‘more personalized’ Siri from Apple at WWDC 25

At Apple’s Worldwide Developers Conference 2025 (WWDC 25), a range of updates to the company’s operating systems, services, and software were unveiled, including a new visual style called “Liquid Glass” and a revised naming scheme. However, the event was notably silent on the much-anticipated AI-powered, more personalized version of Siri initially introduced at WWDC 24.
Craig Federighi, Apple’s Senior Vice President of Software Engineering, briefly addressed the Siri update during the keynote. He mentioned the team is continuing to develop features to enhance Siri’s personalization, but that the project requires more time to meet Apple’s standards. He indicated that more details would be shared over the next year, implying that the launch is postponed until 2026 or later.
Background on the more personalized Siri
The upgraded Siri was originally announced in 2024 as a major advancement for Apple’s virtual assistant. It promised to leverage artificial intelligence to better understand personal context such as relationships, communications, routines, and more. The goal was for Siri to be more helpful by enabling actions within and across various apps.
Despite these ambitions, reports from Bloomberg revealed that the version in development encountered significant technical issues. The assistant only performed as expected about two-thirds of the time, leading Apple to consider it unready for release.
In March 2025, Apple officially delayed the Siri update and made a leadership change—removing John Giannandrea, the SVP of Machine Learning and AI Strategy, from the project and appointing Mike Rockwell, previously Vision Pro’s lead, as the new head for Siri development. This shake-up signals a strategic effort to realign and improve the AI assistant’s progress.
Interim AI strategies and related updates
While the AI-powered Siri remains delayed, Apple has partnered with OpenAI to bridge the capability gap. When users ask questions Siri cannot answer, queries are redirected to ChatGPT. Additionally, the upcoming iOS 26 release will feature improvements to Apple’s AI image generation app, Image Playground, which now integrates ChatGPT technology as well.
At WWDC 25, Apple introduced several other AI enhancements: developers will gain access to on-device foundational AI models; live translation will be available across Messages, FaceTime, and Phone; new features for visual intelligence; an AI-enhanced Workout Buddy for Apple Watch; AI-powered scripting and automation in the Shortcuts app; and AI integration into the Xcode developer environment.
DeepFounder AI Analysis
Why it matters
Apple’s delay in releasing a fully AI-powered and personalized Siri highlights the significant technical challenges even for leading tech companies in delivering AI experiences that meet high-quality standards. For startups and founders, this signals a continuing shift toward AI integration needing robust reliability and user trust before mainstream adoption. The strategic move also underscores the importance of on-device AI capabilities, signaling a trend toward privacy-conscious AI solutions that do not rely exclusively on cloud processing.
Risks & opportunities
The postponement creates risks for Apple as competitors like OpenAI, Google, and Anthropic rapidly advance their AI assistants and services, capturing market mindshare and user expectations. For the broader market, the opportunity lies in building reliable, context-aware AI assistants tailored for specific domains or devices. Startups can capitalize on filling gaps left by large companies struggling with general-purpose AI assistants, by delivering niche or optimized AI solutions for vertical markets, improving performance and user satisfaction.
Startup idea or application
A potential startup could develop an AI assistant platform focused on privacy-preserving, personalized AI that operates primarily on-device but seamlessly integrates cloud fallback for more complex queries. This assistant could be tailored for professional or family contexts, learning specific communication styles and workflows, and enabling cross-app actions with enhanced reliability. Such a product would draw on Apple’s approach but aim to accelerate time to market and flexibility, targeting users desiring smarter yet private assistant technology.
AIAppleSiriWWDCArtificial Intelligence
Visit DeepFounder AI to learn how to start your own startup, validate your idea, and build it from scratch.
📚 Read more articles in our DeepFounder AI blog.