GCP Generative AI Leader Practice Question 35: Multimodal Search Mixed Files
Google Cloud Generative AI Leader · Test 4 · question 5 of 10
An operations team stores its knowledge base across PDFs, spreadsheets, and slide decks, and some of those files hold the answer in a chart or a screenshot rather than in plain text. Leadership wants employees to type one question and get an answer synthesized from whichever file actually contains it, whether that file is mostly text, a table, or an image, without anyone manually tagging each file's content type in advance. Which Google Cloud capability fits?
Show answer and explanation
Answer:
Gemini Enterprise gives employees a single, multimodal search interface that draws on data across the organization and is designed to analyze and retrieve information from documents that combine text, tables, charts, images, and other visual elements, including PDFs, spreadsheets, and slide decks. That directly matches the requirement to answer from whichever content type actually holds the information, with no manual pre-tagging.
Option A could extract text from images with OCR, but it is a one-off preprocessing step for a single file, not an integrated search experience across a mixed knowledge base, and it does nothing for tables or charts. Option B translates language; it has nothing to do with combining text, tables, and image content into one searchable answer. Option C is built to pull fields out of one known document layout, such as an invoice; it does not generalize across PDFs, spreadsheets, and slide decks with varying content.
Reference: What is Gemini Enterprise?