Assignee AI: A Natural Language-Driven Tool for Automated Cloud Infrastructure Operations and Maintenance

Assignee AI is a natural language-driven tool for automated cloud infrastructure operations and maintenance. It addresses the pain points of traditional AWS management methods (cumbersome and error-prone console operations, Terraform requiring a dedicated language, CDK requiring code) and enables stateless, code-free infrastructure management. Users can generate AWS resources by describing their needs in natural language, with features such as cost estimation and manual approval, significantly lowering the barrier to cloud operations.

The popularity of cloud computing brings flexibility to infrastructure management, but it also presents complexity challenges:

• AWS Console: Cumbersome click operations, prone to errors and difficult to reproduce
• Terraform: Requires learning a dedicated language, complex state file management
• AWS CDK: Requires writing code, has startup dependencies

Assignee AI proposes a new approach: directly interact with cloud infrastructure using natural language—users can generate resources by describing their needs without code or dedicated languages.

Core Architecture Design

Assignee AI adopts an AI-native design, built around large language models:

• Natural Language Understanding Layer: Receives descriptions, parses intentions, extracts parameters, and clarifies ambiguous requirements
• Resource Configuration Generation Layer: Generates AWS resource configurations, automatically adds tags, and calculates estimated costs
• Execution and Verification Layer: Generates API calls, supports manual approval, and verifies resource status
• Stateless Design: Queries AWS APIs in real time to obtain status, avoiding state synchronization issues

Technical Implementation Points

• Integrates advanced LLM APIs, ensures stable output through Prompt Engineering
• Encapsulates AWS SDK, supports core services like EC2 and RDS
• Security permissions: IAM role control, least privilege principle, audit logs
• Error handling: Automatically cleans up resources when creation fails to avoid orphaned resources

Key Functional Features

• Intent-driven creation: Describe needs in everyday language (e.g., "a PostgreSQL database with 10GB storage")
• Intelligent tag management: Automatically adds tags for cost center, environment, creator, and project
• Real-time cost estimation: Calculates monthly fees based on AWS pricing API before creation
• Manual approval mechanism: Adheres to Human-in-the-loop, requiring confirmation for each change
• Status query: Queries the status of existing resources using natural language

Application Scenarios and Value

• Rapid setup of development and testing environments, reducing preparation time
• Non-technical users can self-service cloud resources, reducing reliance on operations teams
• Standardizes infrastructure and follows best practices
• Improves cost visibility and avoids unexpected bills

Assignee AI represents the development direction of AI-native operations (AIOps):

• From "Infrastructure as Code" to "Infrastructure as Conversation": Lowering the barrier to use, natural interaction
• From "Declarative" to "Intentional": Users only need to describe their needs, AI decides the implementation path
• From "Automation" to "Intelligence": Understanding context, handling ambiguous needs, providing suggestions

Current Limitations

• Complex multi-resource architectures require manual planning
• Limited support for non-standard configurations
• Dependent on external LLM APIs, with latency and cost implications

Future Directions

• Support more cloud platforms like Azure and GCP
• Introduce local LLMs to reduce latency and costs
• Add resource optimization suggestions (e.g., identifying idle resources)
• Support automatic optimization and tuning of infrastructure

As a capstone project for a generative AI course, Assignee AI demonstrates the innovative application of AI in the cloud operations field. By combining natural language with cloud APIs, it lowers the barrier to infrastructure management and democratizes the use of cloud resources. Although its current functions are limited, its core concept of "natural language conversation with infrastructure" represents the future direction of operations automation, and it is worth the attention of organizations that want to improve efficiency and lower barriers.