Getting Started with GitHub Copilot CLI and Work IQ

Key Concepts

• GitHub Copilot CLI: A command-line interface tool that enables agentic interaction with GitHub Copilot, allowing users to execute tasks, plan, and query data via fixed syntax.
• Work IQ: A plugin for GitHub Copilot that connects the CLI to Microsoft 365 (M365) and Dynamics 365 data.
• Semantic Index: A system that understands the intent and meaning behind natural language queries, moving beyond simple keyword/lexical search.
• Agentic Capabilities: The ability of an AI to autonomously interact with systems, use tools, and make decisions based on user intent.
• Implicit vs. Explicit Memory: Implicit memory refers to the AI’s learned understanding of relationships between data, people, and work patterns; explicit memory refers to user-defined preferences (e.g., "no emojis").
• WinGet: A Windows package manager used to install necessary dependencies.

1. The Role of CLI in the AI Era

The video argues that despite the rise of natural language interfaces (like MCP), Command Line Interfaces (CLIs) remain essential. CLIs provide deterministic interaction—fixed syntax and commands—which reduces the ambiguity inherent in natural language, making them highly reliable for AI agents to execute specific system actions.

2. Work IQ Architecture

Work IQ acts as a bridge between the Copilot CLI and an organization's M365 ecosystem (OneDrive, SharePoint, Outlook, Teams, and Dynamics 365).

• Data Governance: It enforces existing organizational security, including data permissions, sensitivity labels, and Data Loss Prevention (DLP) policies.
• Contextual Awareness: By leveraging the semantic index and memory, the system understands how artifacts, meetings, and people are interconnected, allowing it to provide relevant recommendations without explicit instructions.

3. Prerequisites and Setup

To utilize this integration, the following must be in place:

• Organizational Level: The tenant administrator must onboard and consent to the Work IQ CLI for the organization.
• Individual Level:
  • A GitHub account (with MFA/passkeys recommended).
  • An M365 Copilot license.
• Technical Installation (Windows):
  1. PowerShell: Install the latest version via winget install Microsoft.PowerShell.
  2. Windows Terminal: Install via winget install Microsoft.WindowsTerminal.
  3. Node.js: Required for various CLI operations.
  4. GitHub Copilot CLI: Install via winget install github.copilot.

4. Configuration and Usage

• Terminal Setup: It is recommended to set the new Windows Terminal as the default application and configure PowerShell as the default profile.
• Authentication: Run copilot /login to link the CLI to a GitHub account via a browser-based device code flow.
• Plugin Integration:
  • Install the Work IQ plugin: copilot plugins add work-iq.
  • Accept the EULA: work-iq accept-eula.
• Interaction Modes:
  • Implicit: Asking a question (e.g., "What are my most recent emails?") allows the agent to decide to use the Work IQ tool.
  • Explicit: Using the command work-iq ask forces the agent to query the semantic index directly.
  • Non-interactive: Running commands directly from the shell (e.g., copilot "tell me reasons to not put a superhero base on Mars") for quick, one-off tasks.

5. Notable Commands and Features

• slashskills list: Displays available plugins and their capabilities.
• slashmodels: Allows users to switch between different AI models based on the need for speed versus deep reasoning.
• Ctrl+T: Toggles the "reasoning display" to see the agent's thought process.
• Session Persistence: Users can resume previous sessions using copilot --resume.

6. Synthesis and Conclusion

The integration of GitHub Copilot CLI with Work IQ transforms the command line from a simple text-based tool into a powerful, context-aware agent. By combining the deterministic nature of CLI commands with the semantic intelligence of M365 data, users can perform complex cross-platform tasks—such as retrieving meeting transcripts or identifying urgent work priorities—directly from their terminal. The system is designed to be secure, respecting enterprise data policies while providing a highly personalized experience through implicit and explicit memory.