AI Augmented Architecture Foundations with Sparx EA and GitHub Copilot
AI Augmented Architecture with Sparx EA and GitHub Copilot
AI Augmented Architecture Foundations with GitHub Copilot is Sparx Services’ hands-on training program for enterprise and solution architects who work in the GitHub ecosystem. A five-session curriculum teaches your team to connect GitHub Copilot to AI Power Tools for Sparx EA and apply AI across all four architecture domains. Nine real activities. Your actual repository. No separate AI client to learn.
📍 Available virtual and on-site. On-site travel costs quoted separately.
Your architects are capable. Your tools are ready. So why is so much of the week still consumed by work that doesn’t require an architect?
Your architects are spending expert time on entry-level work.
Hours every week go to transcribing content into EA models, manually tracing dependencies, running conformance checks, and assembling briefings for stakeholders who can’t open Sparx EA. None of this requires architectural expertise. All of it is time your most experienced people could spend on design, facilitation, and judgment.
Mechanical work moves to AI. Architectural work stays with the architect.
The course teaches your architects to use GitHub Copilot Agent mode to handle the repository mechanics: model creation, dependency tracing, conformance checking, and stakeholder documentation. Every architectural decision stays firmly in human hands.
Your team already has Copilot. It isn’t connected to Sparx EA.
GitHub Copilot is in your team’s VS Code environment today. But without the EA MCP Server configured and a structured approach to writing architecture briefs and grounding Copilot with your MDG standards, the Sparx EA repository remains inaccessible to the AI your architects already use.
Connect the tools your team already has to the repository they already work in.
Foundations configures GitHub Copilot Agent mode to work directly with the live Sparx EA repository. No new AI client. No new desktop application. Architects work in VS Code the same way they always have, with the added ability to direct Copilot to do architecture work in Sparx EA.
Architecture demand is growing. Architecture team headcount isn’t.
Every organization is adding systems, changing processes, and facing regulatory requirements that need architectural coverage. The work is increasing. The team isn’t.
AI augmentation scales what your existing team can do.
Architects trained on AI-assisted modeling and analysis can cover more ground without sacrificing quality or governance standards. The capacity freed by automating mechanical tasks is capacity redirected to the design, facilitation, and judgment work only an architect can do.
There’s no clear path from “we should use AI in EA” to “our architects actually do.”
Most architecture teams know AI could help. Few have a structured approach to identifying which tasks to automate first, how to configure the toolchain against their specific MDG standards, or how to maintain appropriate human oversight without losing the productivity benefit.
A structured curriculum from first configuration to confident daily use.
Five sessions take architects from understanding the case for AI augmented architecture, through configuring Copilot Agent mode with the EA MCP Server, to completing nine hands-on activities across all four AI Augmented Architecture domains using GitHub Copilot in the VS Code environment where their team already works.
A Training Program Built for Practicing Architects. In the Environment They Already Work In.
AI Augmented Architecture Foundations with GitHub Copilot is a five-session, 20-hour curriculum designed for architects who already know Sparx EA and already use VS Code. Every session is hands-on. Every activity uses real architecture work. Every architect leaves with a configured, working environment tailored to their repository and their MDG standards — all inside the VS Code workspace they already have.
No New Tools to Learn
This course is built entirely on GitHub Copilot in VS Code — the AI platform your team already has through GitHub Copilot Business or Enterprise. Architecture AI work lives in the same environment as the code and configuration your models describe.
The Architect Stays in Charge
Human-in-the-loop is not a slide. It’s a design principle woven into every activity. Copilot Agent mode makes every tool call visible in the execution trace. Every hands-on task has an explicit review checkpoint. Students leave knowing which oversight model to apply to which task type, and why.
Your Repository. Your MDG. Your Standards.
The course doesn’t use sample data. Architects configure the toolchain — including the copilot-instructions.md custom instructions file — against their own Sparx EA repository and MDG technologies. The grounding they build in Session 2 is the one they use in production.
Five Sessions. Four Domains. Nine Hands-On Activities.
No fluff. No slides that could be an email. Every session builds something architects use on the job.
AI Augmented Architecture Foundations
Phase 1: Learn GitHub CopilotEstablishes the why and the mental model before any tool is touched. Introduces the four AI Augmented Architecture domains (Modeling, Analysis, Governance, and Stakeholder Engagement) and the scenarios each one addresses. Covers the intelligent automation stack for architecture, the 80/20 principle applied to architecture work, and the “Architect Stays in Charge” framework that governs every subsequent activity. Architects complete their automation inventory and send their first GitHub Copilot prompt.
Working with GitHub Copilot for Architecture Work
Phase 1: Learn GitHub CopilotThe full Copilot skill-building session before the EA toolchain is introduced. Covers VS Code workspace setup for architecture work, the copilot-instructions.md custom instructions file as the grounding mechanism, working with Copilot Chat for architecture tasks using @workspace and #file references, Copilot Agent mode for multi-step document workflows, Copilot Edits for multi-file architecture work, and prompt engineering for architects. Architects build their custom instructions file and complete their first real Chat and Agent mode tasks.
AI Power Tools for Sparx EA
Phase 2: Add the EA ToolchainIntroduces AI Power Tools for Sparx EA: what the MCP server is, how the live Sparx EA connection works in VS Code via .vscode/mcp.json, and what the skills layer provides across the four domains. Architects configure the EA MCP Server in VS Code, extend their custom instructions with MDG standards and repository structure, and verify the connection with their first live repository tasks using Copilot Agent mode. The human-in-the-loop framework is applied specifically to EA operations, including reading the Agent mode execution trace to monitor what Copilot is doing in the repository.
Architecture Modeling with Copilot + Sparx EA
Phase 3: Domain Activities: ModelingThe first applied session. Three hands-on modeling activities covering the three primary ingestion sources architects work with, all executed in Copilot Agent mode with the live EA MCP Server connection:
- Activity 1: Model from spreadsheet data: ingest an application inventory and produce stereotyped elements with a diagram
- Activity 2: Model from a requirements document: extract and model architecture elements from a project specification
- Activity 3: Model from data in an external system: query a CMDB, Jira project, or other connected source; create elements; review relationship discovery candidates
Between activities, a focused discussion on reviewing AI-created models: common mistakes, the correction discipline, and how “flag, don’t guess” in the custom instructions changes the output.
Analysis, Governance, Stakeholder Engagement + Build Your Own Skill
Phase 3: Domain Activities: Analysis, Governance, Stakeholder EngagementCompletes the four-domain coverage with six additional activities, all executed in Copilot Agent mode:
- Activity 4: Trace dependencies across a model: full dependency graph traversal; downstream impact identification
- Activity 5: Compare documentation to a model: assess architecture drift between a product specification and the EA repository
- Activity 6: Validate a model for modeling language compliance: conformance audit against MDG standards with severity-classified findings
- Activity 7: Create diagram notes for accessibility: generate plain-language descriptions of diagrams for non-EA audiences
- Activity 8: Create images from diagrams: export diagram images for presentations and stakeholder communications
- Activity 9: Create your own modeling skill: build a reusable skill file in the VS Code workspace for a modeling task you run repeatedly, shareable with the team via Git
Closes with the curriculum path (follow-on specialist courses) and a personal 30-day action plan.
Cohort Pricing
All programs are private. Your architects only. Content is configured to your repository, your MDG standards, and your architecture scenarios.
| Format | Delivery | Investment |
|---|---|---|
| Virtual Cohort ★ Recommended | 5 sessions × 4 hours · delivered via video conference | $1,450 per student |
| On-Site Cohort | 5 sessions × 4 hours · delivered at your location | $1,450 per student · travel costs additional |
Minimum cohort size: 8 students ($11,600)
Fewer Than 8 Architects?
1:1 mentoring programs are available for individual architects or small groups who need the same skills outside the cohort format. Contact us to discuss →
Common Questions
What is AI Augmented Architecture Foundations with GitHub Copilot?
AI Augmented Architecture Foundations with GitHub Copilot is a private, cohort-based training program from Sparx Services that teaches enterprise and solution architects to use GitHub Copilot Agent mode and AI Power Tools for Sparx EA to automate the mechanical parts of architecture work: model creation, dependency analysis, conformance checking, and stakeholder documentation. Human judgment stays at the center of every architectural decision. It is a five-session, 20-hour curriculum designed for architects who already use Sparx EA and work in VS Code.
How does this course differ from the Claude variant?
The course covers the same architecture domains, the same nine hands-on activities, and the same human-in-the-loop principles. The difference is the AI interface: this course is built entirely on GitHub Copilot in VS Code, with grounding via copilot-instructions.md and MCP configuration via .vscode/mcp.json. The Claude variant uses Claude Desktop with Cowork. If your team is already standardized on GitHub Copilot, this course delivers the same architecture AI capabilities inside the tools they already use.
What do participants need before starting?
Participants need to be familiar with Sparx Enterprise Architect — this is not a Sparx EA training course. Participants also need GitHub Copilot Business or Enterprise access and VS Code installed. No prior AI or Copilot experience is required. We will discuss your team’s starting point during the discovery call.
Is AI Power Tools for Sparx EA required?
Yes. The course is built around the AI Power Tools EA MCP Server, which provides the live connection between GitHub Copilot and the Sparx EA repository. Session 3 covers installation and VS Code configuration. Participants need a working Sparx EA instance on their machine. Contact us if you have questions about compatibility with your current EA version.
Do participants use their own repository?
Yes — and this is by design. The most important learning happens when architects run real tasks against their own models, using their own MDG standards. Session 3 specifically covers configuring custom instructions for your repository structure and MDG context. We can also provide a sample repository for participants who prefer not to use a production model during training.
What is the time commitment?
Five sessions of four hours each — 20 hours total. Sessions can be run on consecutive days, spread across a week, or scheduled to fit your team’s calendar. Between sessions, we assign one short real-work task (30 to 60 minutes) to consolidate the prior session before the next one begins.
How many participants can join?
The minimum cohort size is 8 students. There is no strict maximum, but we recommend 12 or fewer to maintain the quality of hands-on facilitation. For larger groups, contact us to discuss split cohorts or additional facilitator support.
What if we have fewer than 8 architects who need training?
1:1 mentoring programs are available for individual architects or small groups. These are structured around your specific use cases and repository, with direct coaching rather than cohort delivery. Contact us to discuss the right format for your situation.
What do participants leave with?
Each participant leaves with: a configured GitHub Copilot and AI Power Tools for Sparx EA environment inside VS Code; a tested copilot-instructions.md custom instructions file grounded with their own MDG standards; a personal architecture prompt library; nine completed hands-on activities from their own repository; and a personal 30-day action plan. Skills built during the course live in the Git repository and are shareable with the whole team.
Is this a virtual or in-person program?
Both options are available. Virtual delivery is the default. On-site delivery is available at an additional travel cost, quoted based on location. Contact us during the discovery call to discuss which format works best for your team.
What courses come after Foundations?
Foundations introduces all four AI Augmented Architecture domains. The Sparx Services specialist curriculum provides full depth in each: separate courses cover Architecture Modeling, Architecture Analysis, Architecture Governance, Building Architecture Dashboards with Power BI, Building Custom MDG Technologies, and Model Validation. Foundations is the prerequisite for all specialist courses. All specialist courses work with either the Claude or GitHub Copilot client.
The Plan · Build · Train Framework
Foundations is the core Train offering for architecture teams who have completed or are completing the Build phase. If your team is still evaluating AI readiness or hasn’t yet deployed AI Power Tools for Sparx EA, start with Plan →.
Ready to have a conversation?
No pitch. A conversation about where your architecture team is and what you’re trying to do.
Book a Discovery Call