Google is shutting down the Gemini feature known as Gems, replacing it with a new system called skills. This change comes as competitors like Meta’s Muse and Instinct push their own all-in-one AI agents into the mainstream. The company confirmed that the custom assistants users built will not be deleted. Instead, they will be automatically converted into skills usable across different AI tasks.
A message inside the Gemini app warns users that the migration begins on 17 November 2026. Google states it will handle the conversion so no manual work is required. The existing Gems remain functional until that date.
Launched in 2024, Gems were designed to teach AI specific jobs without repeating instructions. Pre-made examples included a learning coach, a brainstorming assistant, a career guide, a coding partner, and an editor. Users could also build their own versions for needs like a running coach, nutritionist, or vacation planner. These custom tools could be shared with others, a move intended to boost the Gemini app’s popularity.
The shutdown highlights a pattern in Google’s strategy. The company often gives new AI features distinct brand names and icons only to merge them later. This approach has been a weakness for Google long before the AI boom. At one point, the firm ran multiple different messaging and communication apps simultaneously.
Even after conversion, the former Gems are not as easy to use as typing text into a chatbot. To activate a skill, users must type a forward slash followed by the task name in a thread. This interface suits engineers but is not ideal for regular people.
The news of the shutdown was first reported by 9to5Google over the weekend.
What it means
For people making things, the shift means less friction in the future but more friction now. You will not lose your custom tools, but you will have to wait until late November to use them in the new format. The requirement to type a slash before using a skill adds a step that breaks the flow of conversation. It is a clear signal that Google is prioritising system architecture over immediate user convenience.
