Gemini may let anyone build AI agents on Android

Gemini may let anyone build AI agents on Android. Google Gemini is testing customisable Daily Brief and on-device management of Spark. Building a custom AI agent has so far meant paying for a Workspace or Enterprise plan and doing it through a browser. Google's latest Android testing suggests that might not stay true for much longer. The latest update of the Google app, Gemini's portal in the Android platform, version 17.45.14, has added a "New agent" tab hidden within the sidebar below recent conversations. Currently, this feature is available only to Workspace and Enterprise users via the website, wherein they can write about any custom task. The differentiating factor here is that the mobile version seen in tests offers five template tasks rather than an open-ended one. These suggestions are individual as opposed to corporate; examples include Family for household duties and activities that are done together; Tutor for homework and quiz creation; Group for coordinating activities with a lot of people; Marketing for social media posts and campaign organisation; and Finance for tracking expenses and billing reminders. Whether there will be any opportunity to customise the app once a template has been picked out is not known, or whether the process will be simpler on the phone compared to creating tasks in the desktop version to allow more flexibility. There are also no assurances about access: since this option is so personalised, making it available to all users, even those using Gemini for free makes more sense than just for paid versions of Workspace and Enterprise.