Describe Your Requirements
The conversational path lets you describe your infrastructure needs in plain English. Guardian Pro's AI architect guides you through a structured dialogue, asks clarifying questions, and then generates a production-ready infrastructure template tailored to your answers.
No infrastructure expertise is required. Just describe what your application does, and the Wizard handles the rest.
Starting a Conversation
- Navigate to the Infrastructure Wizard page from the main navigation or the Dashboard.
- On the start screen, select Describe Your Requirements.
- You are taken to the conversation interface, where you can type your first message.
Your opening message should describe what you want to build. You do not need to be technical -- the Wizard understands natural language descriptions and will ask for clarification if it needs more detail.
How the Conversation Works
The AI architect uses a multi-turn conversation to understand your requirements. Here is the typical flow:
Turn 1: Describe Your Application
Start by telling the Wizard what you want to build. Good opening messages include context about:
- What the application does -- "I'm building a real-time analytics dashboard for IoT sensor data"
- Expected scale -- "We expect around 10,000 concurrent users" or "This is a small internal tool"
- Key features -- "It needs a REST API, a PostgreSQL database, and a React frontend"
- Environment tier -- "This is for production" or "I need a dev/staging environment"
You do not need to mention specific AWS services. Saying "I need a database" is perfectly fine -- the Wizard selects the most appropriate service based on your requirements. However, if you have a preference for a specific service, you can mention it directly.
Example opening messages:
"I'm building a SaaS application with a React frontend and a Python API backend. It needs a PostgreSQL database, user authentication, and file upload capabilities. This is for production use with around 500 daily active users."
"I need infrastructure for a data processing pipeline that ingests CSV files from S3, transforms them with Python, and stores results in a data warehouse. It should run on a schedule every hour."
"We're deploying a containerised microservices application with three services. It needs a load balancer, a Redis cache, and a MySQL database. Production tier."
Turn 2: Clarifying Questions
Based on your initial description, the AI architect may ask follow-up questions to fill in gaps. These questions help the Wizard generate more accurate infrastructure. Common questions include:
- What programming language or framework does your backend use?
- Do you need a CDN for your frontend?
- Should the database be multi-AZ for high availability?
- Do you need a CI/CD pipeline included in the template?
- What are your data retention requirements?
Answer these questions naturally. You can provide as much or as little detail as you like -- the Wizard uses sensible defaults for anything you do not specify.
Turn 3: Generation
After gathering enough information (up to three conversation turns), the Wizard generates your infrastructure template. You receive:
- A summary of the architecture it has designed
- A list of all AWS services included and why each was chosen
- The complete infrastructure template ready for review
The conversation is limited to three turns to keep the process focused and efficient. If the Wizard has enough information after your first or second message, it may generate the template early. If it needs more detail, it asks targeted questions to fill the gaps.
Tips for Better Results
Be Specific About Your Workload
The more context you provide, the better the result. Compare these two descriptions:
| Vague | Specific |
|---|---|
| "I need a web app" | "I need a web application with a React frontend, a Node.js API, and a PostgreSQL database for about 1,000 users" |
| "Set up a database" | "I need a MySQL database with 100 GB storage, multi-AZ for high availability, and automated daily backups" |
| "Build me some infrastructure" | "I'm deploying a containerised Python API with three microservices behind a load balancer, with Redis for caching" |
Mention Your Environment Tier
The Wizard generates significantly different templates based on whether you are building for development, staging, or production:
- Development -- Single-AZ, smaller instance sizes, minimal redundancy, cost-optimised
- Production -- Multi-AZ, appropriately sized instances, full redundancy, monitoring, and alerting
If you do not specify a tier, the Wizard defaults to production-grade settings.
State Preferences Explicitly
If you have a strong preference for a specific AWS service, mention it directly:
"I want to use Aurora Serverless for the database, not standard RDS."
"Use Fargate for the containers -- we don't want to manage EC2 instances."
"Include a CloudFront CDN in front of the S3 static site."
The Wizard respects explicit preferences while ensuring the overall architecture remains coherent and well-integrated.
Include Non-Functional Requirements
Non-functional requirements help the Wizard make better design decisions:
- Compliance -- "We need HIPAA compliance" or "SOC 2 is required"
- Performance -- "API response times must be under 200ms"
- Availability -- "99.9% uptime SLA required"
- Budget -- "Keep monthly costs under $500" or "Optimise for cost"
What Happens After Generation
Once the Wizard generates your template, you are taken to the template review screen where you can:
- Inspect every resource that will be created
- Understand the architecture decisions and why each service was chosen
- Make adjustments before deploying
See Review Your Template for details on the review process.
Conversation Management
Starting Over
If the conversation is not heading in the right direction, you can start a new conversation at any time by clicking Start Over in the Wizard interface. This clears the current conversation and returns you to the starting screen.
Conversation History
Your Wizard conversations are saved and can be revisited later. Each conversation is associated with the account you were viewing when you started it. You can find previous conversations and their generated templates on the Wizard page.
Common Scenarios
Here are some common use cases and example descriptions to help you get started:
Web Application
"I'm building a full-stack web application with a React frontend hosted on a CDN, a Python FastAPI backend running in containers, a PostgreSQL database, and user authentication with email/password login. Production tier for approximately 5,000 monthly active users."
Data Pipeline
"I need a data pipeline that processes JSON events from a message queue, transforms them with Python, and loads them into a data warehouse. The pipeline should run automatically when new data arrives. We expect about 1 million events per day."
Static Website
"Deploy a static website built with Next.js. It needs a CDN, a custom domain with HTTPS, and an S3 bucket for hosting. Development environment."
Microservices
"I have four containerised microservices: an API gateway, a user service, an order service, and a notification service. They communicate via message queues. I need a load balancer, a shared PostgreSQL database, and a Redis cache. Production tier with CI/CD."
Next Steps
- Scan a Repository -- Alternatively, let the Wizard detect requirements from your codebase.
- Review Your Template -- Understand what was generated and how to customise it.
- Deploy -- Take your template from review to a running AWS environment.