Google News

Gemini in Chrome: Skills, Auto-Browse, and a Browser-Wide AI Assistant

Gemini in Chrome has been expanding steadily since its initial launch, gaining significant new features in early 2026. On 28 January, Google announced auto-browse and a rebuilt side panel experience powered by Gemini 3. On 14 April, Skills arrived: a way to save and reuse custom AI prompts with a single click. The regional footprint has grown from a US-only desktop product to a multi-platform feature available across six new APAC markets.

What Gemini in Chrome is

Gemini in Chrome is an AI assistant embedded in the browser itself, accessible via a side panel. It is distinct from Google Search, AI Overviews, and AI Mode. It operates on the page you have open, rather than on a search query, and can read the content of your current tab to answer questions or complete tasks.

The distinction matters for publishers. AI Overviews and AI Mode operate inside Search and determine which sources Google cites when answering queries. Gemini in Chrome operates inside the browser and provides answers from the tab you are already on, without generating a citation from Google’s index. The two systems overlap in their potential to reduce outbound clicks, but the mechanism is different.

Auto-browse: 28 January 2026

On 28 January 2026, Google launched auto-browse for AI Pro and Ultra subscribers in the US, powered by Gemini 3. Auto-browse lets users specify a multi-step task, such as comparing product prices across multiple sites, booking an appointment, or filling in forms, and have Gemini complete it without further interaction.

The update also introduced a redesigned side panel that provides contextual assistance based on the page currently open. Users can ask questions about what they are reading, get summaries, or start tasks from within any tab.

Auto-browse positions Chrome as a direct competitor to dedicated AI browser products and to third-party browser extensions that offer similar agentic functionality.

Skills: 14 April 2026

On 14 April 2026, Google launched Skills in Chrome, letting users save frequently used AI prompts as reusable one-click tools. Skills require Chrome 147 or later and English (US) language settings at launch.

A user might create a Skill to extract nutritional information from recipe pages, or to assess how well their CV matches a job listing. Once saved, a Skill appears in the side panel and runs against whatever page is currently open.

Skills are a practical acknowledgement that power users were already constructing similar workflows manually. The formalisation gives Google a structured way to learn about how Gemini is being used across the web.

APAC rollout: April 2026

In April 2026, Gemini in Chrome went live in Australia, Indonesia, Japan, the Philippines, Singapore, South Korea, and Vietnam, available on desktop and iOS (except iOS in Japan). This follows the US launch and marks the product’s first significant expansion outside North America.

Personal Intelligence: coming to Chrome

Personal Intelligence, which allows Gemini to draw on data from Gmail, Google Photos, and other connected apps, launched in AI Mode for AI Pro and Ultra subscribers in January 2026. Google confirmed in January 2026 that the same capability is coming to Gemini in Chrome, though no specific date has been given for that rollout.

What this means for publishers

The SEO implications of Gemini in Chrome are different in kind from those of AI Overviews or AI Mode. AI Overviews affects what users see before clicking into a site. AI Mode affects whether a traditional SERP appears at all. Gemini in Chrome affects what users do once they have already arrived at a page.

For informational content, the risk is that users increasingly ask Gemini to summarise or extract what they want rather than reading directly. For transactional pages, auto-browse introduces the possibility of AI agents completing tasks on behalf of users, which could change how conversion events are measured and attributed.

The most meaningful near-term signal to watch is whether referral traffic patterns begin to diverge between content types: informational pages may see shorter session times, while transactional pages may see changes in how conversions occur if auto-browse starts completing multi-step flows on users’ behalf.

Sources

More news