Home › Certifications › Azure AI Fundamentals AI-901 › AI Workloads and Capabilities
41 multiple-choice questions and 24 flashcards on AI Workloads and Capabilities, about 16% of the Azure AI Fundamentals AI-901 bank. Every one carries a written rationale.
AI Workloads and Capabilities is one of 6 chapters in CoStudy's Azure AI Fundamentals (AI-901) bank, and it holds 41 of the bank's 250 multiple-choice questions — roughly 16% 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.
10 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.
Which capability converts spoken audio into written text?
Answer: C — Speech-to-text in Azure Speech in Foundry Tools
C) Correct — speech-to-text transcribes an audio stream into text. A) Runs the opposite direction. B) Converts text between languages. D) Pulls fields out of scanned forms, not audio.
Which service should you use to extract printed and handwritten text from images and PDFs?
Answer: C — Azure Vision in Foundry Tools, using its read OCR capability
C) Right answer - the read OCR capability of Azure Vision in Foundry Tools returns printed and handwritten text from images and PDF pages. A) Translation converts text that has already been extracted. B) Speech works on audio streams and files. D) Content Safety scores material for harm and extracts nothing.
Sentiment analysis labels text as:
Answer: A — Positive, negative, neutral or mixed
A) Correct — the service returns those sentiment classes with confidence scores at document and sentence level. B) That is a custom image or text classifier. C) That is a spam filter. D) That is a survey coding scheme, not a sentiment output.
Which Azure Vision in Foundry Tools capability returns a natural-language description of an image?
Answer: D — Captioning, which returns a sentence describing the image
D) Right answer - captioning generates a short sentence summarizing what the picture shows. A) OCR returns the characters found, not a description. B) Face detection returns coordinates for faces only. C) Translation changes the language of text and never inspects the image.
Which is NOT a typical capability of Azure Vision in Foundry Tools?
Answer: C — Stock trading signals
C) Right answer - producing trading signals is a financial forecasting task and has nothing to do with image analysis. A) Tagging returns visual concepts present in a picture. B) OCR reads text out of images. D) Captioning returns a sentence describing the scene.
Named entity recognition detects entities such as:
Answer: D — People, organizations, locations and dates in text
D) Correct — named entity recognition tags spans of text and assigns each a category such as person, organization, location or date. A) Belongs to audio processing. B) Is a statistical measure, not an entity. C) Comes from object detection in images.
Which scenario calls for summarization rather than translation?
Answer: A — Turn a 100-page report into a one-page brief
A) Correct — condensing content while staying in the same language is summarization. B) Crosses languages, so it is translation. C) Is speech synthesis. D) Is a vision task.
Which best describes image classification?
Answer: C — Assigning one or more labels to the image as a whole
C) Right answer - classification labels the whole image rather than any region inside it. A) Bounding boxes per object are the output of object detection. B) Pulling characters out of the picture is OCR. D) Making a picture from a description is image generation.
Which task is best suited to the OCR capability of Azure Vision in Foundry Tools?
Answer: B — Extracting printed and handwritten text from a scan
B) Right answer - OCR reads printed and handwritten characters out of images and PDFs and returns them as text. A) Recognizing a car model is image classification or object detection, a different Vision capability. C) Turning speech into another language is handled by Azure Speech in Foundry Tools. D) Forecasting from historical numbers is not a vision workload at all.
Which is NOT a common AI workload category on Azure?
Answer: C — Quantum simulation
C) Right answer - quantum simulation runs on Azure Quantum and is not an AI workload category. A) Computer vision interprets images and video. B) Information extraction pulls structured fields out of documents, forms, images and audio. D) Generative AI produces new text, images or code and is the centre of the current exam.
3 cards from the 24 in this chapter.
What does sentiment analysis produce, and what is its main limitation?
A sentiment label — positive, negative, neutral or mixed — usually with confidence scores, at document and sentence level. Its limitation is tone, not truth: sarcasm, comparisons and mixed statements in one sentence are frequently misread, so it should drive triage rather than final decisions.
What is Artificial Intelligence (AI)?
Systems that perform tasks that typically require human intelligence — perception, reasoning, language, learning.
Sentiment analysis?
Detects positive/negative/neutral tone in text. Scored 0-1.
These are a sample. The full AI Workloads and Capabilities chapter runs 65 items with per-chapter progress tracking, on the web and in the iOS app.
Open Azure AI Fundamentals AI-901 in CoStudy →