Meta launched a new Mac application for Meta AI on Thursday that includes system-wide dictation and screen context analysis using the Muse Spark model. The software allows users to speak commands that the assistant interprets across any open window, similar to recent updates from Google’s Gemini app and existing tools like Wispr Flow. This release forms part of a broader push to integrate Meta AI with business operations, enabling merchants to link Instagram, Facebook, and Google Workspace accounts for automated insights. The assistant now generates reports on campaign performance and audience engagement while also analysing publicly available competitor data. Users can request the creation of proposal decks, draft documents, or build spreadsheets directly through voice or text prompts. During the fiscal second-quarter 2026 earnings call, chief executive Mark Zuckerberg highlighted the opportunity to sell these agents to businesses for automating customer support and inquiries on platforms like WhatsApp and Instagram.
The move shifts the focus from simple chat interfaces to functional agents capable of performing specific tasks within the operating system. By combining voice input with visual context, the tool reduces the friction of interacting with software that traditionally relies on typed commands. This approach aligns with industry trends where generative AI moves beyond generating text to executing actions across different applications.
* Dictation functions globally across all installed applications
* Muse Spark model processes visual screen context
* Integration connects Instagram, Facebook, and Google Workspace accounts
