How to recognise each AI workload from a scenario.
From Ultra Transcenders AI-901 by Tony Rough (publishing soon)
Each workload is defined by what goes in and what comes out. Use the “Recognise it by” column as your test: identify the input (a prompt, a conversation, text, audio, images, documents) and what the system must do with it.
| Workload | Recognise it by | Examples |
|---|---|---|
| Generative AI | Creates new content from a prompt | Marketing emails, newsletter drafts, press releases, product images from descriptions, artwork |
| Agents / conversational AI | Two-way dialogue; interprets questions and replies | Web chat bot for events or refunds, internal help bot, smart home device answering “What will the weather be like today?”, weather chat bot |
| Text analysis (NLP) | Interprets existing written language | Review sentiment, profanity checks in reviews, news monitoring for negative mentions, email work vs personal (text classification), summarising an article |
| Speech | Audio in or out | Captions for video, call transcripts, reading messages aloud |
| Computer vision | Interprets images and video | Photo contains a person, item shape matches, face detection, counting animals in video, flooded areas in aerial photos, brand logos |
| Information extraction | Pulls structured fields from documents, images, audio, video | Medical admission forms, scanned invoices, receipts |
As a rough map to the rest of this book: generative AI models are covered in Chapters 2 and 4, agents in Chapter 5, text analysis in Chapter 6, speech in Chapter 7, computer vision and image generation in Chapter 8, and information extraction in Chapter 9.
This note is one section of Ultra Transcenders AI-901: Microsoft Azure AI Fundamentals, an independent study guide that explains every topic the exam covers by technology, with comparison tables, diagrams and the common traps, plus a glossary linked to Microsoft Learn.
Publishing soon on Amazon in Kindle and paperback editions.
About the book · Free AI-901 glossary · All AI-901 study notes
Fairness, reliability and safety, privacy and security, inclusiveness, transparency and accountability, and how to tell them apart in a scenario.
How guardrails, system messages, grounding and user experience design reduce harm in a generative AI solution.
What happens between a prompt and a response, and how inference differs from training.
Which model setting controls randomness, which controls length and cost, and which ones are not set at deployment.
Agents as model plus instructions, knowledge and tools, and the auto, required and none tool_choice values.
Speech to text, text to speech, translation, batch transcription and speaker recognition compared.
How analyzers turn documents, images, audio and video into structured JSON, and how to call them from code.