1. Generative AI (and Agentic AI)
🤖 Generative AI
Creates new content (text, images, code, etc.)
Typical scenarios:
Chatbots answering questions
Writing emails, blogs, or reports
Code generation
Image generation
âś… Example:
“Generate responses to customer queries using company documents”
➡️ Generative AI (LLM-based chatbot)
🧠Agentic AI
Takes actions autonomously (multi-step reasoning + tools)
Typical scenarios:
AI assistant that books meetings, sends emails, queries systems
Automated workflows with decision-making
âś… Example:
“An AI assistant that reads emails, extracts requests, and schedules meetings”
➡️ Agentic AI
đź“ť 2. Text Analysis (Natural Language Processing)
Works with understanding text, not generating it.
Capabilities:
Sentiment analysis
Key phrase extraction
Named entity recognition
Language detection
âś… Example:
“Determine whether product reviews are positive or negative”
➡️ Sentiment analysis (Text analytics)
🗣️ 3. Speech AI
Works with audio / spoken language
Capabilities:
Speech-to-text
Text-to-speech
Speech translation
âś… Example:
“Convert recorded customer calls into written transcripts”
➡️ Speech-to-text
🖼️ 4. Computer Vision
Works with images and video
Capabilities:
Image classification
Object detection
Face detection
OCR (reading text from images)
Example scenarios:
Detecting defects in manufacturing → Object detection
Identifying objects in photos → Image classification
Reading text from invoices → OCR
đź“„ 5. Information Extraction
Focuses on pulling structured data from unstructured content
👉 Often overlaps with:
Text analysis
Computer vision (OCR + parsing)
Capabilities:
Extract names, dates, totals from documents
Process forms and invoices
Example scenario:
“Extract invoice number, date, and total from scanned invoices”
➡️ Information extraction (Document Intelligence)
⚖️ 6. How to Identify the Correct Workload (Exam Logic)
Use this quick mapping:
Scenario ClueAI WorkloadGenerate new contentGenerative AITake actions / automate tasksAgentic AIUnderstand text meaningText analysisWork with audioSpeechAnalyse images/videoComputer VisionExtract structured data from docsInformation extraction
🧩 7. Common Exam Traps
⚠️ Generation vs Analysis
Writing text → Generative AI
Understanding text → Text analysis
⚠️ OCR vs Information Extraction
Just reading text → Computer Vision (OCR)
Pulling structured fields → Information Extraction
⚠️ Chatbot Types
Simple FAQ bot → Generative AI
Bot that performs actions → Agentic AI
🧠8. Simple Memory Trick
“Create, Understand, See, Hear, Extract, Act”
Create → Generative AI
Understand → Text Analysis
See → Vision
Hear → Speech
Extract → Information Extraction
Act → Agentic AI
âś… Summary
To identify AI workloads in AI-901:
Focus on what the system is doing
Match it to a workload category
Watch for subtle differences (e.g. generate vs analyse)