Home › Certifications › Azure AI Fundamentals AI-901 › Multimodal, Speech, Vision and Content Understanding
48 multiple-choice questions and 18 flashcards on Multimodal, Speech, Vision and Content Understanding, about 19% of the Azure AI Fundamentals AI-901 bank. Every one carries a written rationale.
Multimodal, Speech, Vision and Content Understanding is one of 6 chapters in CoStudy's Azure AI Fundamentals (AI-901) bank, and it holds 48 of the bank's 250 multiple-choice questions — roughly 19% of the total. That proportion is not arbitrary: chapters follow the certifying body's published exam outline, and the number of questions in each is set by that domain's published weight, so the share of your practice time this chapter takes matches the share of the real exam it accounts for.
Studying by chapter is worth doing once you have a diagnostic score. A single overall percentage tells you whether you are close; it does not tell you which domain is dragging. Working a weak chapter in isolation, and re-testing it in isolation, is the fastest way to move a score that has stalled — and it is why the mock exams in CoStudy report by domain rather than as one number.
3 questions drawn from this chapter, with the full rationale shown — the controlling principle behind the right answer, and why each wrong option tempts and fails.
A team submits an image-generation prompt and receives a refusal. Which explanation is MOST likely?
Answer: B — A content safety filter blocked the prompt as violating
B) Correct — content safety filters evaluate image prompts and block requests that violate policy, returning a refusal instead of an image. A) A generation model is not required to be fine-tuned first. C) Image models do not need an accompanying audio file. D) Masks are only needed for edits, not for generation.
Extracting the vendor, date and total from an invoice is best done with which capability?
Answer: B — The prebuilt invoice model in Azure Document Intelligence
B) Right answer - Azure Document Intelligence in Foundry Tools ships prebuilt models for invoices, receipts and similar forms that return named fields directly. A) Speech works on audio, not on a document image. C) Translation converts wording between languages and returns no structured fields. D) Content Safety flags harmful material and does no extraction.
A product team wants a synthesized narrator whose voice matches their brand's existing spokesperson style. Which capability of Azure Speech in Foundry Tools addresses this?
Answer: C — Custom neural voice
C) Correct — custom neural voice lets an organization build a synthetic voice matched to a chosen speaker, subject to consent and responsible-use review. A) Diarization separates speakers in recognition output. B) Language identification detects the spoken language. D) Pronunciation assessment scores a learner's speech.
4 cards from the 18 in this chapter.
How do you give a model visual input in a prompt, and what can it then do?
Include the image alongside your text in the same user message, then ask a question about it. The model can describe the scene, read text in it, compare it to another image, answer questions about its content, and extract values from a photographed form or chart.
You have a mixed pile of PDFs, photographed forms and call recordings and need the same set of fields from all of them. Why does Content Understanding fit?
Because it accepts all of those modalities against one schema, so you define the fields once instead of building an OCR pipeline, a vision pipeline and a transcription-plus-parsing pipeline and reconciling three different output shapes. The schema is the contract; the modality is an input detail.
What is a multimodal model?
A model that accepts more than one kind of input — text plus images, and in some cases audio — in a single prompt and reasons over them together. It removes the need to pre-convert an image or a recording into text before the model can use it.
What is image editing with a generative model, as opposed to image generation?
You supply an existing image and a prompt describing the change, optionally with a mask marking the region to alter, and the model regenerates only that region so the rest of the image is preserved. Generation starts from nothing; editing is conditioned on an image you already have.
These are a sample. The full Multimodal, Speech, Vision and Content Understanding chapter runs 66 items with per-chapter progress tracking, on the web and in the iOS app.
Open Azure AI Fundamentals AI-901 in CoStudy →