Skip to main content
AI DevOps Session

Talk to Your Infrastructure

Forget clicking through cloud consoles or writing boilerplate IaC. With Cloudgeni’s AI DevOps, you just describe what you want and the AI builds it for you.

Natural Language

“Create an S3 bucket with encryption”

Production-Ready Code

Complete IaC with best practices

Real-Time Validation

Validates with your IaC tool

One-Click PRs

Creates a PR ready for review

See It In Action

You: Create an AWS VPC with public and private subnets across 3 AZs

Agent: I'll create a VPC with the following configuration:

- CIDR: 10.0.0.0/16
- 3 public subnets (10.0.1.0/24, 10.0.2.0/24, 10.0.3.0/24)
- 3 private subnets (10.0.11.0/24, 10.0.12.0/24, 10.0.13.0/24)
- NAT Gateway for private subnet internet access
- Internet Gateway for public subnets

[Generates 150+ lines of production-ready IaC]


The Drafting Experience

When you chat with the AI agent, you get a complete workspace:

Chat

Have a conversation. Ask questions. Refine your request. The agent remembers everything.

Diagram

See your infrastructure visualized in real-time as resources are created.

Code

Review every file change with syntax highlighting and diff view.

Policy & Cost

Security findings from Checkov + cost estimates from Infracost. Automatically.

Validation

Validate changes in your actual environment before committing anything.
Pro tip: Queue up follow-up messages while the agent is thinking. It’ll process them in order.

Agent Superpowers

The AI isn’t just generating text—it has real access to your infrastructure:

Cloud Integrations

Connects to AWS, Azure, GCP, and OCI with your credentials. The agent can validate generated code actually works in your environment.
Queries all your cloud resources. Filter by managed vs unmanaged, by provider, by type. Find that EC2 instance someone created through the console last month.
Access to all compliance findings across your infrastructure. Filter by severity, status, framework. “Show me all critical findings” actually works.
Creates branches, commits code, opens PRs. The agent has write access to your repos (with your permission, of course).
Spawns actual pipeline runs. GitHub Actions, Azure DevOps, custom webhooks. See validation output before you merge.

What Can You Ask?

Build New Stuff

  • “Create an ECS cluster with Fargate”
  • “Set up a VPC with Transit Gateway”
  • “Add CloudWatch alarms for my RDS”

Fix Security Issues

  • “Enable encryption on all S3 buckets”
  • “Remove public access from RDS”
  • “Fix the SOC2 compliance findings”

Analyze Infrastructure

  • “What’s my monthly cost by service?”
  • “Show me all unencrypted resources”
  • “List resources without tags”

Import Existing

  • “Import that Lambda someone made”
  • “Bring all EC2 instances into code”
  • “Generate IaC for my Azure VMs”

Supported Formats

Terraform

.tf files with modules, variables, outputs. The works.

Bicep

Native Azure Bicep with proper resource declarations.

Terragrunt

DRY configs for multi-environment setups.

Power User Features

Every session is saved. Click Resume to pick up exactly where you left off—conversation, code changes, everything.

Quick Tips

Always review generated code before creating a PR. The agent is good, but you know your infrastructure best.
Be specific: “Create an S3 bucket” works, but “Create an S3 bucket named app-logs-prod with encryption and 90-day lifecycle” gets you exactly what you need.
Reference existing code: Say “Add a read replica to the RDS in modules/database” to help the agent understand your codebase.

Troubleshooting

ProblemSolution
Session won’t startCheck that you have a repo and cloud account connected
Code generation failsBe more specific or break into smaller requests
Validation errorsCheck the logs—usually a provider config issue
Can’t create PRVerify Git integration has write permissions

Keep Going