Foundations by tool · The deep path

Foundations of AI-Augmented Architecture with GitHub Copilot and Sparx EA

You've got GitHub Copilot in VS Code and Sparx EA — now learn to do AI-Augmented Architecture with them. A private cohort course, taught on your own models and standards, powered by our own AI Power Tools for EA.

Format: private team cohort · Hands-on, on your repository · Prerequisites: Sparx EA + VS Code with GitHub Copilot

Access isn't capability

Turning on Copilot doesn't make a team good at AI-augmented architecture, any more than opening Sparx EA makes someone an architect. The agent amplifies expertise; it doesn't manufacture it. To get reliable results, an architect has to know how to direct the work — what to model, what standards to check, what "good" looks like in their repository.

That's what this course builds. Not button-clicking — the craft of working with GitHub Copilot as a capable collaborator that takes your direction, works your live repository from inside the editor, and handles the mechanical steps, while your judgment stays in the loop.

Program overview

What the cohort covers

The surface

GitHub Copilot in VS Code

Working fluently in the editor your team already uses: prompting Copilot for architecture work, directing multi-step tasks, and keeping control of the outcome — alongside the code and configuration it supports.

The connection

Sparx EA via MCP

How AI Power Tools connects Copilot to your live repository through VS Code's built-in MCP support — elements, packages, diagrams, tagged values — and why operating the running model directly changes everything.

The work

The four use cases

Modeling, analysis, governance, and stakeholder engagement — practiced on your real models against your own MDG standards.

Curriculum

Module by module

1

Orientation & the connection

Wire the EA MCP Server to GitHub Copilot via VS Code's built-in MCP support, against your live Sparx EA repository. Understand the architect-drives/AI-executes model, and run your first guided task end to end without leaving the editor.

2

Modeling on real inputs

Turn spreadsheets, documents, and inventories into properly stereotyped, connected diagrams. Reverse-engineer source data into the repository. Work within your MDG, not a generic default.

3

Analysis & governance

Relationship discovery, impact tracing, and dependency health; then validating against your standards and producing review-ready findings. Where automation checks completeness — and where you confirm correctness.

4

Stakeholder engagement & habit

Produce business-readable outputs from the model, and build the working habits that make this stick after the cohort ends.

Why this beats generic AI training

A generic Copilot course teaches you to autocomplete code. This teaches you to do architecture — on your repository, in your MDG, with the connective tooling we build and use ourselves. Because the EA connection lives inside VS Code, architecture work happens right where your GitHub-standardized team already works. You leave with a portfolio of real work produced during the program, not exercises. And because it's the deep path, you go further than any other tool allows.

Bring the cohort to your team.

A conversation first — we'll confirm prerequisites and scope a cohort to your repository and goals.

Talk to us →