Samuel Boivin | Nurphoto | Getty Images<\/p>\n
Alphabet’s<\/a>\u00a0Google Cloud\u00a0on Thursday introduced the Gemini agent, a single AI agent for work \u2060that can answer questions, handle tasks, create content and write \u200bcode.<\/p>\n The move comes \u200bas tech \u200bcompanies race to sell AI that can work autonomously.<\/p>\n The announcement comes after OpenAI’s September launch of always-on \u2060agents \u200ccalled dots<\/a> that chase user goals \u2060across apps on their own. Earlier last month, Meta<\/a> released Muse<\/a>, a personal AI agent that can shop, book travel, \u200csend emails and make payments for users.<\/p>\n Alphabet shares were up marginally in early trading.<\/p>\n The Gemini agent plans the work, uses tools, connects to a company’s business systems and returns finished \u2060work in documents, email and developer environments, Google said.<\/p>\n The \u200cagent works directly inside Google \u200cWorkspace apps including Gmail, Docs, Sheets and Calendar, and can also be used through Microsoft<\/a> 365 and Slack.<\/p>\n It picks the best model for each task, \u2060currently running on Google’s Gemini models and \u2060Anthropic’s Claude models, with other models to be added later.<\/p>\n Users can create “coworker agents” \u200bthat act as \u200cteam members, with their own email address and access only to the information they are given.<\/p>\n Versions built for financial services and legal work are now in \u200bpreview, and versions for government, \u200chealthcare and retail are coming soon.<\/p>\n