community postAL Developmentcommunity - not Microsoftreviewed - checked by Opus

How Do I: Choose the Right AI Assistant for AL Development?

GitHub Copilot's usage-based billing with model multipliers, starting June 1, prompts a task-based choice of AI tool for AL work. Copilot in VS Code suits code generation that needs workspace context, while Claude.ai or the desktop app gives frontier reasoning models without multipliers for architecture, review and error analysis. The author later noted that this post conflated two billing models and pointed readers to a follow-up correction.

post think about IT

Steven Renders · 2026-05-04 · 980 words · about 4 min

"The right tool depends on the task, not on loyalty to a single subscription."

Key points

  • GitHub Copilot has structural advantage in VS Code due to editor integration and access to workspace symbols and AL objects
  • Claude.ai and Claude desktop app provide frontier reasoning models without Copilot multipliers, making them cheaper for complex analysis and architecture work
  • Pricing multipliers vary significantly by model: Haiku at 0.33x, Sonnet at 1x, and Opus at 27x under Copilot’s usage-based billing
  • Different AI tools suit different workflows: Copilot for inline completions, Claude for reasoning, and Copilot Agent for multi-file refactoring
  • The June 1 pricing change makes the cost of frontier models inside Copilot visible and may justify moving reasoning work to separate tools

Quotes

  • “The right tool depends on the task, not on loyalty to a single subscription.” (Core principle for selecting the most economical and effective AI tool for different development activities)

Context

  • Features: GitHub Copilot in VS Code, Claude.ai, Claude desktop app, Copilot Agent mode, editor integration, usage-based billing, model multipliers, agentic refactoring
  • Versions: June 2026

Source: think about IT - Steven Renders, community blog. Summary, key points and quotes are derived (CONTENT-NOTICE.md); read the original for the full text.