CUPERTINO, 26 September 2025 — Apple is quietly developing an internal “ChatGPT-like” iPhone app, code-named Veritas, intended to assist engineers in testing and refining a major overhaul of Siri. The tool, which is not meant for public release, allows Apple staff to simulate conversational interactions, evaluate contextual behaviour, and stress-test the new assistant’s capabilities ahead of its anticipated launch.
Veritas is designed to support a more fluid, multi-topic chat experience: users can reference past conversations, ask follow-ups, and experience longer exchanges across different topics. It mirrors features seen in popular AI chatbots and gives Apple a sandbox to determine how well the revamped Siri responds to real-world conversational patterns.
The Strategy Behind Veritas
With the new Siri redesign delayed, Apple appears to be leveraging Veritas as a bridge from research to production. The app allows engineers to test hypotheses on user intent, conversational memory, and cross-app actions without rolling out unfinished code to millions of users. According to public signals, Apple is developing multiple sub-systems for the assistant—some powered by internal models, others integrating external models like Google’s Gemini—as it tests architectures and user interaction flows.
Apple’s executives have publicly expressed caution about treating AI as a standalone chatbot. Craig Federighi, Apple’s SVP of Software Engineering, has repeatedly emphasised that the ideal AI experience should be “integrated into everything you do, not a bolt-on chatbot on the side.” The Veritas app may be part of how Apple reconciles that philosophy with the practical demands of conversational AI.
Roadmap, Risks & Investor Signals
The new Siri is internally slated for release in early 2026, with many projecting a debut as part of iOS 26.4. But Apple has faced repeated delays due to reliability and architectural challenges. Veritas is expected to decommission once final versions of Siri are ready for broader user testing.
From an investor lens, Veritas underlines how Apple is accelerating its push into generative AI. The internal app bets heavily on advanced language models, context retention, and app automation—features that potentially elevate Siri from a voice assistant to a full assistant interface across iOS. If Apple succeeds, it could regain competitive footing against rivals like Google Assistant and Alexa.
However, the path is strewn with risks: model performance must match user expectations, data privacy and processing on device vs cloud remain contentious, and the launch must avoid the narrative of a delayed “AI catch-up.” Veritas is Apple’s controlled, risk-mitigated step toward those ambitions.
