Question

A global customer-support platform receives incoming support tickets in dozens of languages. The engineering team wants to automatically detect each ticket's language and translate the body into English before routing it to agents. They have no ML expertise and no labeled training data, and they need production-grade quality with minimal operational overhead. Which approach best meets these requirements?

Accepted Answer

Use the prebuilt Cloud Translation API, which auto-detects the source language and translates text on demand without requiring any model training.

Answer

Train a custom sequence-to-sequence translation model on Vertex AI using the historical ticket corpus and deploy it to an online prediction endpoint.

Answer

Use AutoML Natural Language to build a custom translation model by labeling a subset of tickets with their target-language equivalents.

Answer

Export tickets to BigQuery and use BigQuery ML to build a text-classification model that maps each ticket to an English summary.