Choosing the Stack: Blazor + Azure + Dataverse with AI Input
Why .NET for industrial software reliability, Blazor Server for real-time manufacturing UIs, and Dataverse for multi-tenant data. How GitHub Copilot helped evaluate trade-offs and generate proof-of-concepts.
Series
Building Andon-SSP with AI
This is Part 2 of a 4-part series on building Andon-SSP with GitHub Copilot. Read Part 1 for context on why I chose AI-assisted development.
Coming soon: This post will explore the technology stack decisions behind Andon-SSP—why Blazor Server for real-time manufacturing UIs, why Dataverse for multi-tenant manufacturing data, why SignalR for instant alerts, and how Copilot helped evaluate trade-offs through rapid proof-of-concepts.
What This Post Will Cover
- Why .NET for industrial software: Reliability, performance, enterprise support
- Why Blazor Server over React/Vue: Real-time updates via SignalR, C# end-to-end, reduced complexity
- Why Dataverse for data platform: Built-in multi-tenancy, security, Power Platform integration
- Why SignalR for real-time communication: Sub-2-second alert delivery requirements
- How Copilot helped: Generating POCs comparing SignalR vs WebSockets, evaluating Blazor component patterns, creating entity models
Example: Asking Copilot to Compare Technologies
One of the key decisions was SignalR vs raw WebSockets for real-time production alerts. I prompted Copilot:
"Compare SignalR and WebSockets for real-time manufacturing alerts. Requirements: sub-2-second delivery, automatic reconnection, group subscriptions by production cell, support for 100+ concurrent connections. Generate pros/cons table and recommend approach."
Copilot generated a comprehensive comparison that informed my decision to use SignalR. I'll share the full prompt and response in this post.
Why This Matters for Solo Founders
Choosing the right stack is critical. The wrong choice costs months. AI-assisted technology evaluation lets you:
- Generate proof-of-concepts in hours, not days
- Compare alternatives with real code examples
- Validate architectural patterns before committing
- Reduce risk of costly technology pivots
Part 3 (next): Real examples of Copilot generating production code—entities, SignalR hubs, the Windows Gateway, comprehensive tests.
Questions about technology stack decisions for manufacturing SaaS? Contact me—I'm documenting everything I learned.
💡 About This Blog: AI-Assisted Content Creation
I build software—and write about it—with AI pair programming. Every post on this blog is co-created with GitHub Copilot, not to replace human expertise, but to amplify it.
My Process:
- Prompting: I provide domain knowledge, structure, and strategic direction
- Drafting: Copilot generates content, code examples, and alternatives
- Refining: I edit, validate technical accuracy, and add personal insights
This is the same workflow that built Andon-SSP—an enterprise manufacturing platform shipped faster and better than I could have built solo.
Think AI can't build real products? Let's talk →