Since January 2026, I’ve been handling support tickets in Jira and gradually built a workflow that integrates those tickets directly into my SDF development process.
When I pick up a ticket, I download the XML version of the Jira ticket and add it to my SDF project structure inside a <code>docs</code> folder.
( i have to download, because the jira env is in a secure cloud pc)
Each ticket also gets its own Git branch, for example:
<pre><code>phil/ticket###
</code></pre>
This keeps the development work isolated while still storing the original ticket context alongside the code.
<h2>Working with AI inside VS Code</h2>
I mainly work in VS Code with GitHub Copilot, which makes this workflow extremely convenient.
Because the project files are directly accessible, the AI model can read the context of the ticket and the surrounding SDF files while helping me implement the solution.
Most of the time I use:
<ul>
<li>Claude Sonnet 4.6 for deeper reasoning and implementation</li>
<li>When I run out of tokens, I switch to the free models like GitHub’s Raptor Mini (preview)</li>
</ul>
Even the lighter models can still be useful when they have good context from the project files.
<h2>Copilot Chat Modes I Use</h2>
One thing I really like about Copilot is the simplicity of its built-in chat modes:
Agent:
Runs tasks directly and can create or modify files automatically.
Ask:
Great for questions or feedback on code, but it doesn’t create files.
Plan:
This is more like reasoning together with the AI. It discusses the approach first, and at the end asks if you want to start implementing the plan.
<h2>Adding Context to the Model</h2>
Another useful feature is the “+” button, which lets you attach additional files to the prompt.
This is extremely helpful when you want the AI to understand:
<ul>
<li>the Jira ticket XML</li>
<li>existing scripts</li>
<li>configuration files</li>
<li>related SDF objects</li>
</ul>
The more relevant files you add, the better the output tends to be.
<h2>Features I Haven’t Used Much Yet</h2>
There’s also a “Continue in…” feature that lets you move the conversation to:
<ul>
<li>the CLI</li>
<li>the Cloud</li>
</ul>
I haven’t explored this much yet, but it looks promising for larger workflows.
<h2>One Small Trick That Helps</h2>
To keep things organized, I rename the Copilot chat session to the ticket number.
That way I can easily return to the AI conversation later if I need to revisit the implementation or reasoning.
<h2>Curious About Your Workflow</h2>
This is currently how I use AI together with SDF projects and NetSuite development.
I assume many of you are also experimenting with AI in your NetSuite workflows.
I’d love to see how others are approaching this.
What does your AI-assisted NetSuite workflow look like?

Welcome to the NetSuite SuiteScript Hub!

This channel is your ultimate resource for all things related to NetSuite SuiteScript development. Whether you’re just starting out or you’re a seasoned pro looking to deepen your expertise

NetSuite Suitescript Devs

A developer shares their personal workflow for integrating GitHub Copilot into SDF (SuiteScript Development Framework) and NetSuite development. The process involves downloading Jira tickets as XML, storing them in the project's docs folder, creating per-ticket Git branches, and using VS Code with Copilot's Agent/Ask/Plan chat modes. Key tips include attaching relevant files (ticket XML, scripts, config) for better AI context, using Claude Sonnet 4.6 for complex reasoning with fallback to free models, and renaming chat sessions to ticket numbers for easy reference.

My SDF Workflow with AI (GitHub Copilot)