AI Conductor Academy

Core: 0 of 6 done

Self-paced academy

AI Conductor Academy

Learn how to plan multi-agent workflows and design specialist agents in Grok Bot, at your own pace. The Core path takes about 90 minutes. Optional electives open after M3.

Start with M0 See all modules

Before you start: get Grok Bot

You can finish every lab in this course right in your browser. To try what you learn for real, you’ll need Grok Bot, xAI’s app for working with AI agents. The app is a free download at x.ai/bot; you sign in and link a plan the first time you open it.

Where it runs

  • Mac: Apple silicon or Intel
  • Windows: x64 or Arm64
  • Linux: x64 or Arm64 (.deb, .rpm, or AppImage)
  • Phone: iPhone and Android

What you’ll need

One of these plans, linked when you sign in:

  • Any paid individual Cursor plan, or Cursor Teams
  • SuperGrok, SuperGrok Plus, or SuperGrok Heavy

About the screenshots: every screenshot in this course was taken in the Grok Bot desktop app on a Mac. On other devices, menus and layouts may look a little different.

Checked September 24, 2026, against xAI’s Getting started guide. Plans and platforms change, so check x.ai/bot for the latest.

What you’ll be able to do

By the end of the Core path you’ll be able to do the first five. The electives add the sixth. Every module lists its own objectives at the top, and each module’s main objective is shown through something you do or build. (Outcome 5 is taught in M3.5, just before outcome 4 in M4.)

  1. Match Grok Bot’s four surfaces (agents, channels, computer, routines) to real enablement work. Taught in M1
  2. Configure a specialist agent profile with a name, job, constraints, checkable success criteria, and non-goals. Taught in M2
  3. Specify a multi-agent handoff plan (roles, shared schema, handoff rules, success criteria) that saves work outside chat. Taught in M3
  4. Design a self-paced practice lab (prompt, success criteria, self-check) that an AI agent could build. Taught in M4
  5. Evaluate a Grok Bot workflow for clear outcomes, real evidence, and transfer, including how it should change when the product does. Taught in M3.5
  6. Design an advanced workflow runbook with roles, sequence, gates, and risk controls. Taught in Electives

Big questions this course helps you answer

Core path

M0 · Start hereAbout 7 min

Orientation + Diagnostic

Pick your path. New to Grok Bot? Go straight to M1. Know the basics? Take a 6-question quick check to fast-track.

Start
M1 · Practice labAbout 12 min

Surfaces & Mental Models

Learn Grok Bot’s four surfaces (agents, channels, computer, routines) and when to use each.

Start
M2 · Project 1About 18 min

Specialist Agents

Write a brief a specialist agent can follow, then download it.

Start
M3 · Project 2About 24 min

Orchestration Planning

Plan how agents hand off work and where it’s saved. Finishing M3 opens the electives.

Start
M3.5 · EvaluateAbout 10 min

Evaluate Provided Case

Review someone else’s workflow and suggest how it should adapt when the product changes.

Start
M4 · Project 3About 18 min

Capstone Lite

Design a short practice lab that an AI agent could build for your learners.

Start

Electives open after M3

Elective 1About 30–40 min

Course Review Ops

Run a course review with a comment-approval gate and guardrails for sensitive content.

  • Approval gate
  • Content guardrails
  • Review scope
Open
Elective 2About 30–40 min

Inventory Army & Council

Coordinate agents filling one shared schema, with clear owners and a stop rule for messy data.

  • Shared schema
  • Field owners
  • Stop rule
Open
Elective 3About 30–40 min

Teach-from-Demo

Teach an agent from a live demo while keeping credentials and personal data out.

  • Demo capture
  • Redaction rules
  • Pre-publish gate
Open

Electives open after you finish M3. Passing the M0 quick check doesn’t unlock them on its own.