Beyond Chat: AI That Works, Shops & Understands You

Anthropic’s new agent can access your files, plan tasks, and execute actions on your Mac

Anthropic has launched Claude Co-work, a new product that allows Claude to operate directly on a user’s computer. Designed for non-developers, Co-work can access selected folders, analyze their contents, propose multi-step plans, and execute actions such as organizing files, summarizing documents, identifying duplicates, and managing tasks.

Unlike traditional chat-based assistants, Claude Co-work requests explicit permission before making changes and then performs the work autonomously using system-level commands. The feature currently works inside the Claude desktop app on macOS and is rolling out first to paid plans, with broader availability planned over time.

Claude Co-work represents a major shift from AI as a conversational assistant to AI as an operator. Instead of telling users what to do, it performs real-world actions on their behalf. This signals the beginning of agent-based AI products that integrate directly into operating systems and workflows, reducing friction between intent and execution.

New Gemini feature connects Gmail, Photos, YouTube, Search and more to deliver deeply contextual, personalized AI responses.

Google has launched a major update to its Gemini AI platform called “Personal Intelligence,” a beta feature that lets users connect their personal Google apps to Gemini including Gmail, Google Photos, YouTube history, Search data and more so the AI can give far more contextually rich and personalized answers based on what it knows about their life.

Once enabled in settings, Personal Intelligence lets Gemini reason across multiple data sources at once for example, pulling details from emails, photos and search history to answer questions more accurately or proactively without manually specifying where to look. It does this only if the user opts in, and Google emphasizes that Personal Intelligence is off by default and fully optional.

The feature is currently rolling out in beta for Google AI Pro and AI Ultra subscribers in the United States, and it works across web, Android and iOS platforms. Google plans to expand availability to other countries and eventually to free users over time.

Personal Intelligence marks a significant evolution in mainstream AI assistants from generic knowledge tools to truly personalized assistants that can reason about your life and digital context. Instead of just answering isolated prompts, Gemini can now factor in real user data to deliver tailored advice and proactive help, making interactions faster and more intuitive and reducing the need to repeat background details.

This also shifts AI toward deeper integration within personal ecosystems, where understanding your context not just the world becomes a key competitive advantage. The move underscores how major AI platforms are racing not just on model performance, but on access to and intelligent use of personal data (with user permission).

Major update brings vertical video, 4K upscaling, richer storytelling, and mobile-ready tools across Gemini, Flow, and YouTube Shorts.

Google has rolled out a significant update to its Veo 3.1 video generation model, introducing a suite of improvements that make video creation more expressive, higher quality, and better suited for modern content platforms. The upgraded “Ingredients to Video” feature allows users to generate dynamic clips from multiple reference images with enhanced character and scene consistency even as settings change. For the first time, Veo now supports native vertical video in 9:16, catering directly to TikTok-style and YouTube Shorts formats, and offers state-of-the-art upscaling to 1080p and 4K resolution for sharper visuals. These updates are rolling out across the Gemini app, YouTube Shorts, Flow by Google, Google Vids, the Gemini API, and Vertex AI.

These enhancements mark a step change in AI-generated video quality and usability. By enabling vertical outputs and high-resolution upscaling, Google is aligning its AI video tools with how people actually consume and share content today especially on mobile platforms. The ability to use reference images to maintain continuity in characters and scenes makes storytelling more coherent and creative workflows easier. Moreover, broad availability across multiple platforms means both casual creators and professional filmmakers can tap into powerful generative tools without extra steps.

Search trend analysis gets smarter now with AI-generated suggestions and insights

Google has updated the Google Trends Explore page to include AI features powered by Gemini, its flagship artificial intelligence. The redesigned interface includes a new side panel that automatically identifies and suggests related search terms and helps users compare trends relevant to a topic of interest. The system draws on Gemini’s capabilities to highlight patterns, discover correlated queries, and surface deeper insights right inside the Trends dashboard. This AI-enhanced version also offers suggested prompts and expanded comparison tools, making it easier for journalists, marketers, analysts, and creators to explore data with context and speed. The rollout has begun on desktop and is expanding gradually to more users and regions.

Google Trends has long been a go-to tool for discovering what people around the world are searching for but until now it was essentially raw, human-driven data. With Gemini integration, Trends becomes a proactive AI assistant, synthesizing patterns and suggesting connections that might be hard to spot at a glance. This significantly lowers the barrier for non-specialists to make sense of search behavior, whether for SEO, content ideation, market research, or trend forecasting. Gemini’s integration means data doesn’t just sit there anymore it talks back, helping people ask smarter questions with less manual work.

An open standard designed to let AI agents handle product discovery, checkout, and more transforming how we shop online.

At the 2026 National Retail Federation (NRF) conference, Google introduced the Universal Commerce Protocol (UCP), a new open-source standard that lets artificial intelligence agents navigate the entire e-commerce journey from product discovery and price comparison to checkout and post-purchase support without requiring custom connections for each platform. The protocol was developed in collaboration with major players like Shopify, Etsy, Wayfair, Target and Walmart, and is already backed by over 20 industry partners including payment networks like Visa and Mastercard. UCP will be integrated into AI Mode in Google Search and the Gemini app, enabling U.S. shoppers to buy products directly inside these AI interfaces using payment info stored in Google Wallet, with support for PayPal coming soon.

UCP could fundamentally change how online commerce works by bringing the entire shopping experience into AI agent workflows. Instead of the traditional search → click → checkout journey on retailer websites, AI agents will be able to complete purchases on behalf of users, reducing friction and creating a new point of interaction between brands and customers. For retailers, that means adapting to a world where product visibility and transactions may increasingly happen inside conversational AI experiences, not just on standalone e-commerce sites.