GitHub Copilot has quietly become one of the most helpful tools for anyone working with Business Central. What began as a smart autocomplete has now evolved into a full set of AI-powered features built right into Visual Studio Code. If you spend a lot of time writing AL, refactoring older code, or reviewing complex extensions, Copilot can make your daily work noticeably faster and more enjoyable.

I’ve been using Copilot daily in AL projects over the past months, and it’s changed how I approach both new development and refactoring work.

This post gives you a straightforward introduction to setting up GitHub Copilot in VSCode and walks through the different Copilot modes — Inline Suggestions, Chat, Ask, Edit, Agent, and Plan — and how they fit into an AL developer’s workflow.

---

🚀 Getting GitHub Copilot Set Up

To get started, you’ll need:

• The latest version of Visual Studio Code

• A GitHub Copilot subscription

• VSCode extensions:

  • GitHub Copilot
    

  • GitHub Copilot Chat
    

  • AL Language (from Microsoft)

    

After you install the extensions and sign in with your GitHub account, VSCode will show new Copilot icons in the sidebar and within the editor.

---

🤖 What Copilot Can Do Today

Copilot inside VSCode now comes in several different “modes,” each designed to support a different part of your development workflow:

• Inline Suggestions – autocomplete-style code suggestions as you type

• Chat – a sidebar where you can ask questions, generate objects, or troubleshoot

• Ask – quick explanations based on selected code

• Edit – rewrite or improve code directly in the file

• Agent – perform multi-step changes or workspace-wide refactors

• Plan – shows a preview of changes Copilot wants to make

Here’s how each of them helps when you’re working in AL.

---

✨ Inline Suggestions

(Copilot’s fastest and simplest way to help)

As you type AL code, Copilot offers suggestions similar to autocomplete — but much more capable. Often it completes entire procedures, validates fields, writes triggers, or builds out repetitive patterns.

It’s especially useful for:

• creating standard AL boilerplate

• writing test procedures

• building event subscribers

• generating loops, queries, or simple business logic

---

💬 Copilot Chat

(Your conversation partner while building AL extensions)

Copilot Chat lives in a sidebar panel. You can ask it to:

• generate a full AL object

• explain what code does

• help troubleshoot compilation errors

• propose refactors

• write test codeunits

• outline architecture or design choices

For example:

“Create a page extension for Customer Card that adds a field ‘Loyalty Score’ under the General group.”

Chat will generate the object with fields, triggers, captions, ApplicationArea, and a decent layout.

---

❓ Copilot Ask

(Perfect for quick explanations)

When you just want to understand a piece of code, highlight it and use Ask Copilot.

Good for:

• understanding unfamiliar or legacy code

• reviewing PR changes

• learning AL patterns

• checking what a trigger or event subscriber actually does

Ask is lightweight and fast — ideal when you don’t need the full Chat experience.

---

✏️ Copilot Edit

(Directly rewrite code in your file)

With Copilot Edit, you select a block of code, choose Edit with Copilot, and describe what you want.

Examples:

“Rewrite this to use guard clauses.”
“Extract the business logic into a separate codeunit.”
“Make this more readable and add comments.”

The output appears as a proposed code change you can accept or revise.
It’s one of the most practical ways to improve or modernize AL code.

---

🤖 Copilot Agent

(Copilot that performs multi-step actions)

Agent is where Copilot begins to feel like a collaborator rather than a suggestion engine. You can ask it to perform larger, workspace-wide changes such as:

• applying ApplicationArea properties everywhere

• adding telemetry hooks to selected procedures

• adjusting naming conventions

• updating object references across files

• generating missing XLF translations

Agent performs actions step by step and shows you what it plans to change before applying anything.

---

🗺 Copilot Plan

(See what Copilot wants to do before it does it)

Plan gives you a preview when Copilot intends to make bigger changes. It outlines:

• which files will be touched

• what will be added or modified

• how the changes are structured

For AL projects — where object IDs, dependencies, and specific patterns matter — this is incredibly helpful.

Example:

“Plan how to refactor this codeunit to use interfaces and dependency injection.”

You get a clear step-by-step outline before applying anything.

---

🧭 Quick Guide: When to Use What

---

🔚 Wrapping Up

GitHub Copilot has grown far beyond simple autocomplete. With the combination of Chat, Ask, Edit, Agent, and Plan, it now supports AL developers through nearly every step of building and maintaining Business Central extensions.

Used well, it saves time, reduces repetition, and helps you maintain higher-quality code — without taking away your control or expertise.

Further reading:

• Getting started with GitHub Copilot
• GitHub Copilot · AI coding built your way
• GitHub Copilot · Plans & pricing
• Asking GitHub Copilot questions in GitHub – GitHub Enterprise Cloud Docs

This post is intentionally introductory. In the next articles, I’ll dive deeper into practical AL prompts, real refactoring examples, and common pitfalls when using Copilot in Business Central projects.

==> Chat, Ask, or Edit? Using GitHub Copilot Effectively in AL Development – think about IT