Inspiration
As a Solution Architect, I often found myself spending more time designing solutions than actually building them. Every project started with understanding business requirements, deciding on the right architecture, comparing cloud services, selecting the appropriate technology stack, and estimating infrastructure costs. These activities could take days or even weeks and often involved multiple iterations with stakeholders.
Creating detailed architecture and data flow diagrams was another time-consuming task, especially when ensuring they accurately reflected the proposed solution. Cost estimation added another layer of complexity—determining the right compute resources, choosing suitable pricing tiers, and balancing the trade-offs between a cost-effective MVP and a scalable production-grade deployment. Once the design was finalized, the infrastructure still had to be provisioned manually or through Infrastructure-as-Code, followed by continuous monitoring of costs, permissions, and unused resources to maintain an optimized cloud environment.
I wanted to build an intelligent platform that could automate this entire journey—from a simple business idea to a deployable cloud solution. That's how ArchForge AI was born: an AI-powered Solution Architect that dramatically reduces the effort required to design, validate, estimate, and generate production-ready cloud architectures.
What it does
ArchForge AI transforms a natural language business requirement into a production-ready cloud solution using a team of specialized AI agents. The platform begins by understanding the user's requirements and asking intelligent follow-up questions to gather any missing functional and non-functional details.
Once the requirements are complete, it generates a layered solution architecture along with interactive architecture and data flow diagrams. After user approval, it recommends the most suitable Azure services and pricing tiers based on workload, scalability, availability, and cost optimization. Finally, it generates deployment-ready Infrastructure-as-Code using standardized Bicep templates and YAML configuration files, packaged into downloadable artifacts.
The vision extends beyond deployment. ArchForge AI aims to continuously monitor deployed environments, identify unused resources, recommend cost optimizations, detect security and permission issues, and provide architecture improvement suggestions throughout the application's lifecycle.
How we built it
ArchForge AI follows a modular agentic architecture where each AI agent has a dedicated responsibility. The backend is built using FastAPI for API management and Azure Durable Functions to orchestrate the multi-stage workflow. Azure OpenAI powers the specialized agents responsible for requirement analysis, clarification, architecture generation, diagram generation, pricing recommendations, and Infrastructure-as-Code generation.
Artifacts such as diagrams and deployment templates are stored in Azure Blob Storage, while Mermaid is used to generate architecture and data flow diagrams. The frontend is designed as an Architecture Studio, allowing users to review, approve, and iterate through every stage of the solution before generating deployable artifacts. The modular architecture makes it easy to extend the platform with additional cloud providers and governance capabilities.
Challenges we ran into
One of the biggest challenges was designing an AI workflow that behaves like an experienced solution architect instead of a generic chatbot. We had to ensure that every decision was explainable and validated before progressing to the next stage.
Another challenge was maintaining context across multiple user interactions while orchestrating several specialized AI agents. Structuring architecture outputs into machine-readable formats that could later be converted into diagrams, pricing recommendations, and Infrastructure-as-Code required careful planning. Balancing AI-generated recommendations with deterministic engineering logic—especially for cloud pricing and infrastructure generation—was essential to produce reliable, production-ready outputs.
Accomplishments that we're proud of
We're proud of creating a platform that automates one of the most time-consuming aspects of cloud engineering. Instead of simply generating text responses, ArchForge AI produces tangible engineering artifacts including architecture diagrams, data flow diagrams, Azure service recommendations, pricing estimates, and Infrastructure-as-Code.
The agent-based workflow provides transparency by allowing users to validate every stage before proceeding, making the system practical for real-world engineering teams. We also designed the platform with extensibility in mind, enabling future support for additional cloud providers, governance policies, deployment pipelines, and operational monitoring.
What we learned
Building ArchForge AI reinforced that effective AI systems are not just about using powerful language models—they require thoughtful orchestration, structured outputs, deterministic workflows, and human approval at critical decision points.
We learned the importance of separating reasoning from execution by assigning specialized responsibilities to different AI agents. We also gained valuable insights into combining AI-driven decision-making with cloud engineering best practices, resulting in a system that is both intelligent and dependable.
What's next for ArchForge AI
Our vision is to evolve ArchForge AI into a complete AI-powered cloud engineering platform. Future enhancements include support for AWS, Google Cloud, and Kubernetes, Terraform generation alongside Bicep, CI/CD pipeline generation, security and compliance validation, architecture optimization recommendations, and automatic deployment through GitHub Actions and Azure DevOps.
Beyond deployment, we plan to introduce an AI Operations layer capable of continuously monitoring cloud environments, identifying cost optimization opportunities, detecting configuration drift, recommending security improvements, and proactively suggesting architectural enhancements. Ultimately, we envision ArchForge AI becoming an autonomous AI Solution Architect that supports organizations throughout the entire cloud application lifecycle—from idea and design to deployment, optimization, and ongoing operations.
Log in or sign up for Devpost to join the conversation.