Apple is quietly building a powerful AI infrastructure, not for the public just yet, but for the very people who create its products. While the rollout of Apple Intelligence and a revamped Siri faces challenges, the company is deeply immersed in internal experimentation, deploying sophisticated AI tools to its workforce.
Two new applications are at the heart of this initiative, designed to not only test the boundaries of artificial intelligence but to seamlessly integrate it into the daily workflows of Apple employees. These aren’t simple trials; they’re a strategic effort to refine AI capabilities and directly contribute to the evolution of Apple Intelligence.
The first, called Enchanté, functions as a secure, internal equivalent to ChatGPT. Employees are leveraging it for brainstorming, refining code, polishing documents, and accessing a vast reservoir of knowledge – all within the protective walls of Apple’s security protocols.
Unlike many companies that restrict AI use due to data security concerns, Enchanté is built from the ground up with privacy in mind. It operates exclusively on Apple-approved models, running locally or on private servers, ensuring sensitive information never leaves the company’s control. Access extends to leading models like Claude and Gemini, alongside Apple’s own Foundation Models.
This heightened security unlocks powerful features. Employees can upload documents, images, and files for analysis, and the app can even draw information directly from files stored on their Macs. It’s a comprehensive AI assistant tailored to Apple’s unique needs.
An internal memo highlights Enchanté’s dual purpose: a testing ground for AI and a practical tool for everyday tasks. The app is populated with Apple’s internal documentation and guidelines, making it invaluable across all departments – from engineering and design to marketing and executive leadership.
Launched around November, Enchanté incorporates a robust feedback system, allowing employees to rate the quality of responses. Crucially, it also enables side-by-side comparisons between Apple’s models and those from third-party providers, driving continuous improvement.
Complementing Enchanté is Enterprise Assistant, a more focused AI knowledge hub. Built entirely on Apple’s internal large language models, it serves as a centralized repository of corporate information.
Imagine a single source for answers to questions about executive roles, company conduct, health insurance, vacation policies, or even technical instructions like configuring a VPN on an iPhone. Enterprise Assistant delivers precisely that, streamlining access to vital internal resources.
Like Enchanté, Enterprise Assistant prioritizes feedback, allowing employees to evaluate responses and contribute to the ongoing refinement of the system. This constant stream of data is invaluable for improving accuracy and usefulness.
While the direct connection between employee interactions and future consumer features remains undisclosed, it’s logical to assume Apple is leveraging this internal data to enhance Apple Intelligence. The app’s guidelines explicitly state its purpose is to “support team members across all departments.”
This isn’t an isolated effort. Reports last year revealed Apple’s internal “Veritas” app, another AI chatbot, and ongoing exploration of AI tools for AppleCare agents. These initiatives demonstrate a sustained commitment to AI development.
Ultimately, Apple’s strategy is clear: refine its language models through real-world application within its own workforce. This allows for continuous improvement without exposing consumers to unfinished technology. Despite the challenges in launching Apple Intelligence, a dedicated effort is underway to unlock the full potential of AI within Apple Park.
