Product Launch2 min read

SpaceXAI Launches Grok Bot, AI Agents for Office Work

By , Senior AI ConsultantPublished

SpaceXAI (the merged xAI and SpaceX) launched Grok Bot, a beta AI agent that logs into your work apps and finishes multi-step tasks on its own, starting at $120 a month per person.

SpaceXAI, the company that used to be xAI before Elon Musk merged it into SpaceX earlier this year, has launched Grok Bot. It is an AI agent that can log into the apps and websites you already use and finish multi-step jobs on its own, only checking back in when the work is done or something needs your sign-off.

Think of it less like a chatbot and more like a new hire with its own laptop. You give it an instruction in plain language, it goes off and works using its own cloud computer, and it comes back to you when it is finished. You can run several of these bots at once, and they can message each other to stay in sync when their work overlaps.

This is not a first. OpenAI already sells ChatGPT Work, Anthropic has Claude Cowork, and Microsoft runs Copilot Tasks. Every large AI company has now built a version of the same pitch: stop typing prompts, start handing off entire jobs. Grok Bot is SpaceXAI's attempt to catch up in a race that is already crowded.

What makes this launch worth watching is not the bot itself, it is where it came from. SpaceXAI recently paid 60 billion dollars to buy Cursor, the popular AI coding tool, in the largest startup acquisition on record. Grok Bot is the first product to come out of that combination, and it is currently locked behind Cursor and SuperGrok subscriptions rather than sold on its own, which tells you this is as much about selling more subscriptions as it is about the technology.

Pricing runs from 120 dollars a month for a team seat up to 300 dollars a month for individual access. That is a real cost for a beta product, and it is aimed at people who are already paying for premium AI tools, not first time buyers.

The part business owners should actually worry about is access. A bot that can sign into your email, your files, and your business software is only as safe as the permissions you hand it. Security researchers have already flagged that agents like this can be hijacked by hidden instructions buried in a webpage or an email, a trick called prompt injection. A similar tool from Cursor previously caused an incident where it uploaded entire code repositories without anyone asking it to.

None of this means the idea is bad. Handing routine, multi-step work to software that does not need constant supervision is genuinely useful, and every major AI company is now building toward it. But this is a beta product from a young, expensive, and still-forming combination of companies. If you are curious, treat it as a small supervised pilot on low-stakes tasks, not something you connect to your accounting system or your customer records on day one.


STAY INFORMED

Get AI intelligence like this delivered to your inbox.

Free forever · Unsubscribe anytime


You May Also Find Valuable