AI is transforming many industries, but not all are benefiting equally. Software development adopted AI agents quickly, and the secret was something called Git. This tool provided a backbone for collaboration, making it safe and manageable to work with AI-driven changes. Other fields are now trying to catch up, but they face a big challenge: without a similar infrastructure, AI can cause more chaos than progress.

The Role of Git in Software Engineering

Long before AI was a thing, software engineers relied on version control systems like Git. These tools turned chaotic code changes into a structured process. Each change gets a branch, and every branch creates a diff—an exact record of what was altered. Changes are reviewed, merged, and recorded, ensuring everyone stays on the same page. This infrastructure made collaboration smooth and reliable, even with complex projects.

Thanks to Git, developers can trace every decision, revert to previous versions, and know exactly what changed and why. It enforces discipline—requiring explicit approvals and maintaining a durable history. This setup turned AI from a potential risk into a powerful tool for automation and innovation in software projects.

Challenges in Other Fields Without a Common Backbone

Fields like marketing, architecture, and finance are starting to see the value of AI, but they don’t have a Git-like system in place. Take marketing: an AI might update a competitive intelligence report. But if it makes a mistake or overwrites a section, where is the record of that change? Without a diff or audit trail, it’s hard to see what was altered or to revert to an earlier version.

Similarly, in architecture and urban planning, teams want AI to update models or explore different zoning options. But without a versioning system for spatial data, every change is opaque. An untraceable modification can lead to unreviewable and unmanageable results. In finance, AI can help update models and draft reports, yet without a way to track dependencies and approve changes, automation risks becoming a source of errors rather than a solution.

These examples show why having a structured process is essential. Without a backbone that enforces change control, accountability, and reproducibility, AI can introduce more instability than value. The analogy to Git highlights the need for a shared protocol that all disciplines can adapt to, even if it’s not literal code they’re managing.

Building a Shared Operational Foundation

What these fields need isn’t a direct copy of Git, but a similar system that provides a canonical record of work. This includes fine-grained versioning of artifacts, structured workflows for proposing and reviewing changes, and clear audit trails. It’s about creating a shared contract that ensures everyone understands what has changed, why, and by whom.

For example, a team working on a zoning simulation could use a versioning protocol that tracks every adjustment. When an AI suggests a new zoning plan, the change is recorded with details about what was modified and why. Human reviewers can easily evaluate, approve, or reject the change, maintaining control over the process.

Similarly, in finance, models and assumptions could be versioned and documented. This ensures that updates are transparent and reversible if needed. Structured workflows mean that AI-driven changes are reviewed and approved systematically, reducing risks and increasing trust.

In essence, these disciplines need to adopt a collaborative infrastructure that enforces discipline and transparency. Such a system doesn’t have to be exactly like Git but should deliver the same core benefits: traceability, accountability, reproducibility, and controlled collaboration. Only then can AI truly become a reliable partner across different fields, rather than a source of chaos.

Inspired by

• https://www.infoworld.com/article/4112896/for-agentic-ai-other-disciplines-need-their-own-git.html