Google turns prompts into 'skills'

Google is turning good prompts into saved tools, starting with a new Skills library in Chrome that runs Gemini actions with one click. (blog.google) Google said the Chrome rollout started on April 14 for Gemini in Chrome on desktop. Saved Skills can be called up with a slash command or the plus button, run on the page you are viewing, and use up to 10 shared tabs. (blog.google; support.google.com) A Skill is basically a stored prompt with instructions attached, like saving a template email and turning it into a button. Google’s examples include summarizing product reviews, comparing pages across tabs, and pulling key points from dense articles. (blog.google) Google is also widening the memory and context around Gemini itself. Personal Intelligence, introduced in the United States in January 2026, connects Gemini to apps like Gmail, Google Photos, YouTube and Search so answers can draw on a user’s own data if that user opts in. (blog.google) By March, Google had expanded Personal Intelligence in the United States across Search AI Mode, the Gemini app and Gemini in Chrome. Google said the feature securely connects information across Google apps and lets users choose which apps to link. (blog.google) Google’s help pages say the Gemini web app now runs in more than 70 languages across more than 230 countries and territories, while the mobile app is available in more than 150 countries. That wider footprint gives Google a much larger base for features that depend on saved prompts, connected apps and repeat use. (support.google.com; support.google.com) The company is also pushing beyond one-shot chats toward software that keeps working after the first prompt. Google Workspace Studio, announced in December 2025, lets employees design, manage and share Gemini-powered agents that automate multi-step work inside Google Workspace. (workspace.google.com) Outside Google’s official announcements, Android Authority reported on April 15 that testing builds of Gemini Enterprise show an Agent tab and a Tasks dashboard with sections for Goals, Agents, Connected apps and Files. The publication said the interface points to more persistent workflows inside Google’s work products, though Google has not publicly announced that specific layout. (androidauthority.com) Google has been adding adjacent controls for automation as well. A Gemini help page describes scheduled actions that can rerun instructions later, and a recent Chrome post said auto browse pauses for confirmation before purchases or social posts. (support.google.com; blog.google) Taken together, those pieces move Gemini from answering questions to storing procedures: a prompt becomes a reusable Skill, a connected account becomes context, and an agent becomes something closer to a standing assignment. Google is not just selling chat replies anymore; it is building a library of repeatable actions around them. (blog.google; blog.google; workspace.google.com)