Apple Brings an AI Assistant to Pages, Keynote, and Numbers
- ai-news
- apple
- productivity
- google-deepmind

Apple brings an AI assistant to its office apps
Apple has launched a Creator Studio update that brings a new AI assistant to three of its familiar productivity apps: Pages (word processing), Keynote (presentations), and Numbers (spreadsheets). The assistant, built on both Apple's own models and OpenAI's, understands document context and can perform concrete tasks:
- In Keynote: designing, animating, and refining slides on request.
- In Pages: drafting, rewriting, and restructuring written content.
- In Numbers: calculating, visualizing, and organizing data, including suggesting chart types and building custom tables.
The detail that really matters for businesses is support for the Model Context Protocol (MCP) - an open standard that lets third-party AI agents (not just Apple's own) plug directly into these apps. In other words, Apple is opening its productivity suite to a broader ecosystem of digital assistants rather than keeping everything locked inside its own tools.
The feature is rolling out in beta to Apple Intelligence-compatible devices running iOS 27, iPadOS 27, or macOS 27 and later. The broader update also includes a new "Generate Fill" tool in Pixelmator Pro for editing images from a text description, and an AI assistant in Logic Pro that can set up a vocal track from a single prompt. All of these upgrades will be available free to existing subscribers starting in November.
Why does this matter for small and medium businesses? Because these are tools already used every day - client proposals, presentations, budget tracking - and they're now getting a layer of automation that saves time without requiring anyone to learn a new system.
Source: The Verge
Google unveils Genie 4, a new generation of world simulation
Google DeepMind released Genie 4 as an early research preview - a model that generates more photorealistic and interactive worlds than ever, complete with real-time audio (music, ambient sound, and dialogue) matched to the generated scene. Users can type prompts mid-simulation to change things like the weather or camera angle, and the model now maintains scene consistency for about two minutes, double its predecessor.
Proposed commercial uses include gaming, interactive learning, and robotics simulation - but early testers are also pointing to a more immediate use: rapid prototyping and storyboarding for film projects, well before full production begins.
Source: TechCrunch
Briefly: Firefox improves automatic form filling
Mozilla updated Firefox's Smart Window feature so it can fill out web forms using information pulled from your open tabs - for example, grabbing details from a LinkedIn profile to auto-complete a job application. You can also now ask the browser to "watch" a specific webpage and get notified when something changes on it, such as a sale or a new job listing. The feature is rolling out in beta in the US and Canada.
Source: The Verge
What do these stories have in common? AI is no longer confined to separate chatbots - it's moving directly into the tools we use every day: documents, presentations, browsers. That's exactly the direction we believe in at RealOffice.AI: AI agents that fit into the existing workflow of small and medium businesses, instead of requiring them to learn a new system from scratch.