Apple Scales Back AI Health Coach Ambitions

Bloomberg reported on February 5, 2026 that Apple scaled back an internal effort, code-named Mulberry, which had been intended as a virtual AI health coach; Bloomberg's reporting says the initiative was wound down and that Apple "now plans to take some of the features it had planned for the artificial intelligence-powered offering and roll them out individually over time within its Health app." Coverage of WWDC and watchOS 27 shows Apple introduced an upgraded Siri AI experience and incremental health-tracking and coaching features for Apple Watch, but did not announce a standalone, subscription-style AI health coach (Mashable; CNET).

Editorial analysis - technical context: Public reporting frames the Mulberry effort as an integrated service concept that would have combined device telemetry, longitudinal Health data, and generative AI to produce personalized guidance. Observers writing about the WWDC previews note that the company instead opted for narrower, on-device enhancements to heart-rate tracking, workout guidance, and a more capable Siri on-wrist, which are less ambitious integration points than a full conversational health coach (CNET; AppleInsider; Mashable).

Editorial analysis: Industry coverage places Apple's move in two concurrent trends. First, several consumer-wearable vendors have shipped coach-like features and subscription add-ons, examples in press coverage include Oura Advisor and Whoop Coach, producing what some writers call "AI health coach fatigue" among users. Second, reporting highlights practical limitations: clinical-grade recommendations require high data fidelity, explainability, and regulatory caution. The Verge column characterized the lack of a big-bang health coach announcement as welcome, arguing that many existing coach products oversell capabilities and risk user confusion or false reassurance.

Editorial analysis: Observers will track whether Apple actually surfaces Mulberry-derived features via iterative Health app updates, how tightly those features remain on-device versus cloud-assisted, and whether Apple publishes safety or clinical-evidence summaries supporting any advice. Also relevant are regulatory boundaries: reporting including The New York Times noted that Apple's Siri AI rollout faces regional availability constraints such as Europe, which bear on how health-related features can be deployed and audited.

Editorial analysis: For product teams and ML engineers, the episode underscores two practical constraints when building consumer-facing health AI: (1) integrating heterogeneous device telemetry with longitudinal health records without creating invalid correlations is technically and product-wise difficult, and (2) providing prescriptive or diagnostic guidance elevates legal and regulatory risk and typically demands clinical validation pipelines and transparent uncertainty estimates. Companies in this space often decompose large coach ideas into incremental, verifiable features that can be measured in controlled rollouts, a pattern reflected in the reported approach to roll Mulberry functionality into existing Health features (Bloomberg).

Summary conclusion: Reporting from Bloomberg, Mashable, CNET, and The Verge converges on a pragmatic story: Apple did not ship a full conversational AI health coach at WWDC, Bloomberg reported the Mulberry project was wound down, and industry commentary frames that restraint as responding to technical, safety, and regulatory friction rather than product timidity.