This talk will explore lessons learned from developing the BWX Generative Language Engine, an award-winning Generative AI tool for translation. This engine innovatively employs Generative AI to make contextual decisions in real-time, revolutionizing traditional translation methods. It shifts the paradigm from simple prompt-response mechanisms to dynamic, linguistic problem-solving. This engine has significantly reduced editing time, while maintaining control for linguists and enhancing the human translation experience. The presentation will explore the software architectural patterns and development methodologies employed, highlighting how Generative AI is utilized for real-time, context-sensitive decision-making features.