GCP Generative AI Leader Practice Question 45: AutoML Custom Model No Code
Google Cloud Generative AI Leader · Test 5 · question 5 of 10
A retailer's merchandising analyst has a labeled spreadsheet of past promotions and wants a model that predicts which future promotions will sell out. She knows the business problem well but has never written a line of machine learning code, and the retailer has no data scientist available this quarter.
Which Gemini Enterprise Agent Platform capability lets her build this custom model herself?
Show answer and explanation
Answer:
AutoML on Gemini Enterprise Agent Platform is built for exactly this situation: it automates data preparation, model selection, hyperparameter tuning, and deployment, so someone who understands the business problem but has no machine learning background can train a high-quality tabular model through a console workflow rather than code.
Option B is wrong because Agent Search builds search and RAG experiences over documents; it does not train a predictive model from labeled tabular data. Option C is wrong because Google AI Studio is for experimenting with prompts against an existing foundation model, not for training a new model on the analyst's own labeled records. Option D is wrong because custom training is the option for teams that do write their own training code in a framework such as TensorFlow, which is the opposite of what the analyst needs.
Reference: AutoML training overview