As artificial intelligence reshapes live media workflows, the industry is entering a new phase - shifting from AI captioning to real-time translation. This paper investigates whether the performance trajectory of the measurement system - Number Translation Recognition (NTR) evaluations - is on track to follow the same path of rapid improvement seen in Number Edition Recognition (NER) used to measure the accuracy of AI captioning. This paper also proposes a framework for evaluating industry readiness to adopt AI-powered live translation as a production-standard solution. This framework considers technical performance benchmarks, workflow integration, and operational preparedness - providing practical guidance for broadcasters, content owners, and live event producers seeking to expand reach through multilingual delivery.