3 Ways Spacelift Enables You to Use AI for Infrastructure

AI is changing how teams write code, design systems, and ship software. But when it comes to infrastructure, most teams are still stuck in a world of HCL, pull requests, and manual cleanup.

Spacelift Intent changes that. By combining natural language, provider APIs, and Spacelift’s governance and state engine, Intent lets teams use AI to provision, manage, and correct infrastructure safely without sacrificing control.

Here are three practical ways Spacelift lets you use AI for infrastructure today:

Infrastructure as code is powerful, but it isn’t always the fastest or most appropriate tool — especially when you just want to experiment, spin something up temporarily, and tear it all back down again. With Spacelift Intent, you can provision and delete infrastructure simply by describing what you want in natural language.

Instead of writing Terraform, debugging HCL, and waiting for reviews, you can say things like:

• “Create an S3 bucket in us-east-1 with public access blocked.”
• “Spin up a small EC2 instance for testing.”
• “Delete the resources I just created.”

Intent takes that natural language description, provisions the resources directly via provider APIs, and tracks everything behind the scenes. You don’t have to manage configurations or set up state files, but you still get centralized state, history, and auditability.

This is perfect for experimentation. Developers can try new services or spin up test environments without clicking around in the cloud console (and creating drift that you’ll have to deal with later). Platform teams still get what they need: every action is tracked, compliant with policies, and visible. Even deletions are handled safely. High-risk actions require explicit confirmation, so nothing disappears without a human verifying exactly what’s about to happen.

The result is faster experimentation without sacrificing governance. Your developers don’t have to wait to try out ideas, and platform teams can finally say “yes” without worrying about what gets left behind.

Not sure? Here’s a quick video that walks you through how to use Spacelift Intent to accelerate safely:

Most cloud environments have a hidden problem: unmanaged resources. These are the instances, buckets, and databases that were created manually, spun up by scripts, or left behind after a quick experiment. Over time, they become a source of surprise costs, misconfigurations, and compliance gaps. They’re also notoriously painful to bring back under management.

Until now. 

With Spacelift Intent, you don’t have to recreate resources or manually import them through complex workflows; you can use a single natural language command to scan your cloud environment and import existing resources into Intent. Once imported, those resources are tracked, governed, and visible alongside everything else you manage.

That means you finally get:

• A complete inventory of what actually exists
• Historical context for resources that were previously “in the shadows”
• The ability to apply policies and guardrails consistently
  

AI isn’t just creating new infrastructure here; it’s helping you regain control over what you already have, without disruptive migrations or risky rebuilds.

Sounds complicated? We created a demo video that shows how to scan your cloud, import unmanaged resources, and bring them under Intent’s control. 

Drift is inevitable. Someone changes a setting to fix an issue. A resource is resized during an incident. A manual tweak never makes it back into code. With Spacelift Intent, AI helps you find drift that you may have missed previously.

Intent compares the active state it’s tracking with the real state in your cloud provider. Using natural language, you can ask Intent to check for drift, and it will immediately identify discrepancies: for example, when an EC2 instance has been resized outside of Intent.

Instead of discovering these changes weeks later (or during a billing review), you get instant visibility into:

• What changed
• Where it changed
• When it was last refreshed

From there, teams can decide whether to accept the change or remediate it. And they get the same auditability and governance that’s applied to everything else Intent does. This video shows you how it works in practice.

AI isn’t replacing best practices here; it’s helping to enforce them without requiring humans to constantly watch dashboards or reconcile configs by hand.

Spacelift Intent isn’t about letting AI run wild in your cloud or turning your developers into platform engineers (or vice versa). It’s about using AI in scenarios where it makes sense: removing friction, speeding up workflows, and detecting problems that can occur. And it does all this while maintaining the governance, policy enforcement, and audit trails you rely on.

Whether you’re provisioning infrastructure without IaC, bringing unmanaged resources back under control, or detecting drift before it becomes a problem, Intent demonstrates what AI for infrastructure can look like when it’s built for real teams and real environments. Infrastructure doesn’t have to be slow, risky, or invisible. With Spacelift Intent, it can finally be both conversational and controlled.