Elon Musk launches Grok Bot, AI agent with its own computer that does things for you
SpaceXAI and Cursor have launched Grok Bot in beta, a new AI agent designed to handle tasks across apps and websites instead of simply answering
SpaceXAI and Cursor have launched Grok Bot in beta, a new AI agent designed to handle tasks across apps and websites instead of simply answering questions. The companies describe it as an “AI teammate” that can work on tasks independently and return when human approval is needed. Grok Bot gives each AI agent its own computer in the cloud, allowing it to sign into tools and services used by the user. This means the agent can work across inboxes, websites and business applications, including services that may not have a dedicated API.
Read Full Story What can Grok Bot do According to SpaceXAI, Bots can handle multi-step tasks such as updating customer records, processing invoices and working on software bugs. The agent can use tools, follow a sequence of steps and recover from errors instead of stopping after a failed action. Users can also run multiple Bots at the same time. The company says its teams are using separate agents for areas including sales, operations, recruiting and engineering. Bots can also exchange information and hand work between themselves.
Another feature allows users to show a Bot how they complete a particular task. The agent can remember the workflow and use it again later, reducing the need to repeat the same instructions. Grok Bot availability Grok Bot is currently available in beta to SuperGrok Heavy, Cursor Ultra and Cursor Teams Premium subscribers on desktop and iOS. Enterprise customers can join a waitlist for access. The service is also being offered on Windows, while Android support is listed as coming soon. The launch comes as AI companies increasingly move beyond chatbots towards AI agents that can perform tasks on a user's behalf.
Grok Bot is taking that idea a step further by giving the agent access to its own computer and the ability to work across multiple applications. For now, the product remains in beta, so questions around reliability, security and how much control users should give an AI agent over their accounts and applications are likely to remain important as it develops. Ends
The BriefWire