Transforming Software Architecture Design With Intelligent Assistants-A Comparative Analysis

As artificial intelligence tools such as GitHub Copilot, Chat Generative Pre-Trained Transformer (ChatGPT), or Black Box Artificial Intelligence (BlackBoxAI) become more embedded in software development workflows, they are reshaping how architects approach modern software design. While enterprises are increasingly leveraging Generative Artificial Intelligence (GenAI) tools to enhance developer productivity, limited research has been conducted on their impact on software architects. Traditionally, architects utilize various tools to create architectural blueprints, including Unified Modelling Language (UML) diagrams, class diagrams, sequence diagrams, use case diagrams, and state diagrams. Many of these diagrams follow repetitive structures that can be automated with well-defined contexts. This paper examines the role of three AI tools in assisting architects during the design phase, assessing how closely the generated outputs adhere to established architectural principles. By analyzing deviations, the study explores how refined prompts and additional context can improve accuracy. Research also includes cross tool comparison. Ultimately, this research aims to evaluate the potential of AI assistant tools to enhance productivity, ensure design consistency, and support architectural decision-making, while providing best practices for integrating AI into modern software architecture.