When people first encounter AI coding tools like GitHub Copilot, Claude Code, or Cursor, they often focus on the wrong metric: “Can it write perfect code in one shot?” From my experience working with these tools in Business Central development, I’ve found that the real game-changer isn’t one-shot perfection -...
If you pick the same model in two different tools, you won’t get the same experience. Claude Sonnet 4 is the one example: in Claude Code (Anthropic’s native experience) you get large context and explicit control over extended thinking; in GitHub Copilot you get a managed, IDE‑first product with smaller context,...
Introduction Imagine a world where you can hand off a GitHub issue to an AI agent, and it autonomously develops the code, fixes the bug, and creates a pull request for you to review. This is the power of the GitHub Copilot agent, which can be activated directly from a...
Problem When working with agentic code tools like GitHub Copilot, one crucial way to guide the tool is ensuring that the application actually compiles without errors from code analyzers. This validates the syntax and provides immediate feedback on code quality. This blog post focuses on implementing a make file and...
Natural language interfaces for Business Central data access are now possible via Azure OpenAI. This post reviews a prototype repo, CopilotAllTablesAndFields, that exposes Business Central tables through three AOAI functions. The goal is to prove technical feasibility, not deliver a turnkey product.