AI agents improve over time. Prompts get refined, workflows become smarter, and new capabilities are added as your product evolves. But experimenting directly on a live agent can introduce risk.

A small change in prompts or logic can unintentionally affect behavior and disrupt the experience users already depend on.

GTWY solves this with Agent Versions.

The Version feature allows you to create a new iteration of your AI agent without modifying the original one. Your production agent continues running exactly as before, while the new version becomes a safe environment for experimentation and improvements.

This means you can innovate freely without putting your live system at risk.

Creating a new version in GTWY is simple. When you select Create New Version, the platform instantly clones your existing agent. This duplicate version becomes an isolated environment where you can modify prompts, update workflows, integrate new tools, or test different logic.

While you experiment with the new version, the original agent remains stable and continues serving users.

Once you are satisfied with the changes, you can decide whether to adopt the new version, keep both versions for different use cases, or continue iterating further.

This structured approach creates a clear evolution path for your AI systems.

As agents evolve, it becomes important to keep track of why each version exists.

GTWY allows you to add a version description to explain what the version is designed for and how it differs from previous iterations. This makes it easy to understand the purpose of a version when revisiting it later.

In addition to descriptions, GTWY provides a Notes section where you can document how different versions should be used. For example, one version might be optimized for customer support queries while another focuses on technical troubleshooting or internal workflows.

These details help teams manage multiple versions efficiently and maintain clarity as agents evolve.

AI systems improve through experimentation. However, experimentation should never compromise stability.

GTWY Agent Versions allow teams to test new ideas, refine behavior, and introduce improvements without affecting production agents. You can experiment confidently, compare performance between versions, and always keep a reliable fallback available.

This makes it possible to move fast while maintaining the reliability required for real-world applications.

What are GTWY Agent Versions?
Agent Versions allow you to create new iterations of an AI agent without modifying the original one.

Why should AI agents use versioning?
Versioning allows teams to experiment with prompts, logic, and workflows without risking disruptions to the live agent.

What are version descriptions used for?
Version descriptions explain the purpose of a specific version and how it differs from earlier versions.

What are Notes used for in GTWY Versions?
Notes help document how different versions should be used, making it easier to manage multiple agent variations.