Event Moment: Dona Sarkar on Why Coding Agents Are Leading the AI Era

In this moment, excerpted from her keynote, Dona Sarkar, Chief Troublemaker, Enterprise AI Advocacy, Microsoft, explained why coding agents have become the most successful early example of practical AI agents.

Key Takeaways

• Coding Has a Structural Advantage: Software development already provides ideal conditions for AI: centralized repositories like GitHub, consistently updated data, and clearly defined workflows that agents can understand and operate within.
• Built-In Validation Makes AI Reliable: Code has an objective test of success — it either builds and runs or it fails. Compilers and runtimes act as automatic human-in-the-loop validation systems, allowing coding agents to improve rapidly with immediate feedback.
• Why Other Domains Are Harder: Tasks such as finance, tax, or business analysis lack instant verification. Determining whether an output is correct is far more ambiguous, which slows AI agent adoption outside software engineering.