SpaceXAI has published a series of official guides for Grok Bot, which have been written by the internal teams both using and developing the Bot. Eric Zakariasson, an engineer on the team, posted the brand new link recently.
The guides page presents a series of official guides on xAI’s website, which are summed up in a single sentence: Practical playbooks for AI teammates.
Grok Bot is xAI’s early-beta system of AI teammates that sign into a user’s tools, work on their own computer, and come back when they need approval. The Guides hub is a gathering of best practices and processes from the team inside xAI who are already running those teammates for various capabilities.
Five Playbooks, Five Paths
As of this week, the hub has five official guides published on it.
The guides go over topics stretching from “How I run multiple teams of Grok Bots”, to “Grok Bot for mobile app development”, to the most interesting, “Designing Grok Bot with Grok Bot”. These guides are written by both xAI and Cursor’s engineers who are working on Grok Bot, and provide an excellent way for new users to get started.
These initial guides walk users through some of the best use cases for Grok Bot, and how they’re being actively implemented at SpaceXAI. For those who ask the question: “Why do I need Grok Bot?”, this is where you’ll find the best answers.
Whether it be a Chief of Staff for running your business or developing applications from scratch using Cursor Agents in conjunction with Grok Bot, there’s a lot to learn here, even for experienced users who have been using Grok Bot since early access.
Jumping from Guides to Production
The most useful part is that the guides show how to turn a job into a repeatable workflow. The mobile development walkthrough starts with a written role, then adds connected accounts, timed routines, recorded skills, handoffs, and a clear approval boundary.
Its analytics Bot pulls campaign and product data on schedule, sends findings to specialist Bots, and stops before anything spends money. The GTM guide uses the same structure for meeting prep: connect the daily tools, set a morning routine, and request a short briefing built from records. It also shows how to demonstrate a task on the Bot’s computer, save the recording as a skill, and improve that skill with feedback.
The project-management examples become more concrete once several Bots are involved and need to share the same work. One guide pairs a channel with Notion Projects and Tasks databases, while a manager Bot creates the workspace and staffs it from an existing roster. Work moves through a board, and a Bot marks a task blocked before asking the user for missing input. The design guide takes another useful route: give an agent the real production asset instead of a screenshot, then let it build a local playground for direct tuning. Together, those examples give readers templates for defining ownership, preserving human approvals, and deciding when a successful one-off process should become a reusable skill or scheduled routine.
