Google has announced the discontinuation of "Gems" in Gemini, a feature that allowed users to build custom AI assistants for specific tasks. This change comes as all-in-one AI agents like Meta’s Muse and Instinct gain traction. Importantly, the effort users invested in creating these Gems will not be lost; they will be automatically migrated to "skills" that can be utilized across various AI tasks.
Details regarding this transition are being shared within the Gemini app, where users are informed that Gems will evolve into skills starting on November 17, 2026. The company confirms that it will handle the migration to the new format, eliminating any need for user intervention. Until then, the existing Gems will remain fully functional.
Launched in 2024, Gems were designed to enable users to train their AI to perform certain tasks without repeating instructions. Google’s premade Gems included a learning coach, a brainstorming assistant, a career guide, a coding partner, and an editor. Users could also create custom Gems for personal needs, such as a running coach or a vacation planner, and these could be shared, a feature Google hoped would boost Gemini’s popularity.
The sunsetting of Gems underscores a recurring challenge for Google: its tendency to hastily brand and prominently feature every new AI capability within its apps, only to later re-shuffle or merge them. This pattern of inconsistent product strategy predates the current AI era.
Despite the transition to skills, the user experience for these former Gems may still not be as consumer-friendly as simply typing text into a chatbot like Meta’s Muse. Google notes that users will need to enter a forward slash "/" in a task thread to select the desired skill—a user interface (UI) approach typically preferred by engineers over general users.
[AgentUpdate Depth Analysis] Google's pivot from "Gems" to "skills" within Gemini signifies a strategic consolidation in the burgeoning AI Agent ecosystem. While Gems offered customizability, their isolated nature likely hindered widespread adoption. This move aligns with a broader industry trend towards more integrated, interoperable agent capabilities, contrasting with the more modular, tool-centric approach of frameworks like LangChain. By automatically migrating existing user creations and focusing on cross-task utility, Google aims to simplify agent creation and deployment, drawing parallels with OpenAI's GPTs but emphasizing seamless integration. For the future of AI Agents, this shift implies a greater focus on core platform capabilities that allow agents to leverage a shared pool of #skills, rather than being siloed. This could foster a more robust and efficient AI agent marketplace, where reusable skills become foundational components for building increasingly sophisticated and collaborative autonomous agents, ultimately pushing the boundaries of general AI accessibility.
