Domain 3: AI/ML Services
CLF-C02 Exam Domain 3, Task 3.7 | 34% of Scored Content
Learning Objectives
By the end of this section, you will be able to:
- Identify AWS artificial intelligence and machine learning services (Domain 3, Task 3.7)
- Understand AI/ML service categories and use cases
- Compare different AI/ML services for specific requirements
- Understand when to use SageMaker AI vs purpose-built AI services
Overview: AWS AI/ML Services Landscape
AWS provides a comprehensive set of artificial intelligence and machine learning services that enable you to add intelligence to your applications without requiring deep expertise in ML algorithms.
AI/ML Service Categories
| Category | Services | Purpose |
|---|---|---|
| ML Platforms | SageMaker AI | Build, train, deploy custom ML models |
| Computer Vision | Rekognition | Image and video analysis |
| NLP - Text Analysis | Comprehend | Extract insights from text |
| NLP - Speech | Transcribe, Polly, Translate | Speech-to-text, text-to-speech, translation |
| Document Analysis | Textract | Extract data from documents |
| Conversational AI | Lex | Build chatbots |
| Search | Kendra | Intelligent enterprise search |
| Forecasting | Forecast | Time-series predictions |
AI/ML Services for CLF-C02
Note: For the CLF-C02 exam, you need to understand what each service does and when to use it. You do NOT need to know implementation details.
1. Amazon SageMaker AI
Overview
Amazon SageMaker AI is a fully managed machine learning service that enables data scientists and developers to build, train, and deploy machine learning models quickly.
Key Capabilities
| Capability | Description |
|---|---|
| Data Labeling | Ground truth labeling for training data |
| Feature Engineering | Prepare, transform data |
| Bias Detection | Detect statistical bias in data and models |
| AutoML | Automatically build models |
| Training | Managed training infrastructure |
| Tuning | Hyperparameter optimization |
| Hosting | Deploy models with auto scaling |
| Monitoring | Model performance monitoring |
| Workflows | ML orchestration pipelines |
SageMaker AI Components
┌─────────────────────────────────────────────────────────────┐
│ Amazon SageMaker AI │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │
│ │ Label │ │ Build │ │ Train │ │ Deploy │ │
│ │ Data │ │ Models │ │ Models │ │ Models │ │
│ └──────────┘ └──────────┘ └──────────┘ └──────────┘ │
│ │
│ Ground Truth → Studio → Notebooks → Endpoints │
│ │
└─────────────────────────────────────────────────────────────┘SageMaker AI Use Cases
| Use Case | SageMaker Feature |
|---|---|
| Custom ML model development | Studio Notebooks |
| Automated model building | SageMaker Autopilot |
| Large-scale training | Managed training jobs |
| Real-time predictions | SageMaker Endpoints |
| Batch predictions | Batch Transform |
| MLOps automation | SageMaker Pipelines |
SageMaker AI vs Other AI Services
| Scenario | Use SageMaker AI | Use Purpose-Built Service |
|---|---|---|
| Custom ML model | ✅ Yes | ❌ No |
| Image analysis | Possible, but use Rekognition | ✅ Amazon Rekognition |
| Text-to-speech | Possible, but use Polly | ✅ Amazon Polly |
| Chatbot | Possible, but use Lex | ✅ Amazon Lex |
Key Features for Exam
- Fully Managed: No infrastructure to provision
- Pay-As-You-Go: Pay for compute during training/inference
- Integrated: Data prep, training, deployment in one service
- Scalable: From small to massive ML workloads
- Notebooks: Jupyter notebooks for interactive development
2. Amazon Rekognition
Overview
Amazon Rekognition makes it easy to add image and video analysis to your applications using proven, highly scalable, deep learning technology.
Key Capabilities
| Capability | Description |
|---|---|
| Object Detection | Identify objects (cars, pets, furniture) |
| Scene Recognition | Identify scenes (beach, mountain, city) |
| Face Analysis | Detect faces, emotions, age range |
| Face Comparison | Verify if two faces match |
| Text Detection | Extract text from images (OCR) |
| Celebrity Recognition | Identify celebrities in images |
| Unsafe Content | Detect inappropriate content |
| Custom Labels | Train custom models for your objects |
Rekognition Use Cases
| Use Case | Description |
|---|---|
| Content Moderation | Detect inappropriate images/videos |
| Face Verification | User authentication (ID verification) |
| Sentiment Analysis | Detect emotions from faces |
| Document Processing | Extract text from documents |
| Asset Management | Search images by content |
| Security | Detect people in restricted areas |
How Rekognition Works
┌─────────┐ ┌──────────────┐ ┌──────────┐
│ Image │ ───▶ │ Amazon │ ───▶ │ Labels │
│/Video │ │ Rekognition │ │/Metadata │
└─────────┘ └──────────────┘ └──────────┘
│
├── Objects: ["Person", "Car", "Tree"]
├── Confidence: [98%, 95%, 87%]
└── Emotions: ["Happy", "Neutral"]Storage Integration
- S3: Processes images/videos stored in S3
- Asynchronous: Publishes results to Amazon SNS topic when complete
Exam Tips
- Rekognition = Image and Video Analysis
- Uses deep learning for image recognition
- Integrates with S3 for storage
- Uses SNS for async notifications
- Can detect objects, scenes, faces, text, celebrities
3. Amazon Transcribe
Overview
Amazon Transcribe is an automatic speech recognition (ASR) service that makes it easy for developers to add speech-to-text capabilities to their applications.
Key Characteristics
| Feature | Description |
|---|---|
| Speech-to-Text | Convert audio to text |
| Deep Learning | Advanced ASR technology |
| Multiple Languages | Support for global languages |
| Real-Time | Live transcription |
| Batch | Process recorded audio files |
| Speaker Identification | Distinguish between speakers |
| Automatic Punctuation | Adds punctuation automatically |
| Custom Vocabulary | Add domain-specific terms |
Transcribe Use Cases
| Use Case | Description |
|---|---|
| Call Transcription | Customer support calls |
| Meeting Notes | Automatic meeting transcription |
| Captioning | Video subtitles |
| Voice Analytics | Analyze customer conversations |
| Documentation | Medical/legal dictation |
Input/Output Formats
| Input | Output |
|---|---|
| Audio files (WAV, MP3, MP4) | Text transcripts |
| Live audio streams | Real-time text |
| Phone recordings | Timestamped transcripts |
Exam Tips
- Transcribe = Speech to Text (Audio → Text)
- Uses Automatic Speech Recognition (ASR)
- Can identify different speakers
- Supports real-time and batch processing
4. Amazon Polly
Overview
Amazon Polly is a Text-to-Speech (TTS) service that turns text into lifelike speech.
Key Characteristics
| Feature | Description |
|---|---|
| Text-to-Speech | Convert text to audio |
| Lifelike Voices | Natural-sounding speech |
| Multiple Languages | Support for global languages |
| SSML Support | Speech Synthesis Markup Language |
| Neural TTS | Advanced deep learning voices |
| Voice Customization | Adjust pitch, rate, volume |
| Newscaster Style | News reading style |
Polly Use Cases
| Use Case | Description |
|---|---|
| Accessibility | Visual impairment assistance |
| Voice Assistants | Alexa-style applications |
| E-learning | Course narration |
| Audiobooks | Text-to-audio conversion |
| Gaming | Character voices |
| IVR Systems | Interactive voice response |
Polly vs Transcribe
| Service | Direction | Use Case |
|---|---|---|
| Polly | Text → Speech | Give your app a voice |
| Transcribe | Speech → Text | Convert speech to text |
Exam Tips
- Polly = Text to Speech (Text → Audio)
- Uses deep learning for natural speech
- Neural TTS = most realistic voices
- Supports SSML for speech control
5. Amazon Translate
Overview
Amazon Translate is a neural machine translation service that delivers fast, high-quality, and affordable language translation.
Key Characteristics
| Feature | Description |
|---|---|
| Neural ML | Deep learning models |
| Language Pairs | 75+ language pairs |
| Batch Translation | Translate large volumes |
| Real-Time | Live translation |
| Custom Terminology | Domain-specific vocabulary |
| Formality | Formal/informal tones |
Translate Use Cases
| Use Case | Description |
|---|---|
| Website Localization | Multi-language websites |
| Document Translation | Translate documents |
| Customer Support | Multi-language support |
| Content Distribution | Global content reach |
| Communication | Cross-language communication |
Exam Tips
- Translate = Language Translation
- Uses neural machine translation
- Supports 75+ language pairs
- Delivers fast, high-quality translation
6. Amazon Textract
Overview
Amazon Textract automatically extracts printed text, handwriting, and data from any document.
Key Capabilities
| Capability | Description |
|---|---|
| OCR | Optical Character Recognition |
| Handwriting | Recognizes printed and handwritten text |
| Forms Data | Extract key-value pairs from forms |
| Table Data | Extract table structures |
| Document Analysis | Understands document context |
| Identity Documents | Extract from IDs, passports |
| Invoices/Receipts | Understand business documents |
| Relationships | Identifies relationships in data |
Textract vs OCR
| Feature | Traditional OCR | Amazon Textract |
|---|---|---|
| Text Only | ✅ | ✅ |
| Handwriting | ❌ | ✅ |
| Form Fields | ❌ | ✅ |
| Tables | ❌ | ✅ |
| Context Understanding | ❌ | ✅ |
| Document Type Awareness | ❌ | ✅ |
Textract Use Cases
| Use Case | Description |
|---|---|
| Invoice Processing | Extract line items, totals |
| Form Automation | Process application forms |
| Document Search | Search within scanned PDFs |
| Compliance | KYC, identity verification |
| Data Entry | Automate manual data entry |
| Receipt Processing | Expense management |
Supported Document Formats
| Format | Examples |
|---|---|
| Images | PNG, JPG, TIFF |
| PDFs | Scanned and native PDF |
| Forms | Tax forms, applications |
| Tables | Financial statements, reports |
Exam Tips
- Textract = Document Text and Data Extraction
- Goes beyond traditional OCR
- Extracts forms, tables, handwriting
- Understands document context (e.g., knows what to extract from receipts)
- Uses AI/ML for intelligent extraction
7. Amazon Comprehend
Overview
Amazon Comprehend is a natural language processing (NLP) service that uses machine learning to uncover insights in text.
Key Capabilities
| Capability | Description |
|---|---|
| Sentiment Analysis | Detect positive, negative, neutral, mixed |
| Key Phrase Extraction | Find important phrases |
| Entity Recognition | Identify people, places, dates, quantities |
| Topic Modeling | Discover topics in document collections |
| Language Detection | Identify document language |
| PII Detection | Detect personally identifiable information |
| Syntax Analysis | Parse sentence structure |
Comprehend Use Cases
| Use Case | Description |
|---|---|
| Customer Feedback Analysis | Understand customer sentiment |
| Document Categorization | Auto-categorize documents |
| Social Media Monitoring | Track brand sentiment |
| Compliance | Detect PII in documents |
| Knowledge Discovery | Find topics in large document sets |
| Content Moderation | Detect inappropriate content |
Comprehend vs Textract
| Service | Input | Output |
|---|---|---|
| Comprehend | Text (already digital) | Insights (sentiment, entities, topics) |
| Textract | Scanned documents/images | Extracted text + data |
Exam Tips
- Comprehend = NLP Service for Text Analysis
- Natural Language Processing (NLP)
- Extracts sentiment, entities, key phrases, topics
- Works on text data (not images like Textract)
- Can detect PII for compliance
8. Amazon Lex
Overview
Amazon Lex is a service for building conversational interfaces into any application using voice and text.
Key Capabilities
| Capability | Description |
|---|---|
| Chatbots | Build conversational bots |
| Voice & Text | Both modalities supported |
| ASR & TTS | Speech recognition and synthesis |
| Intent Recognition | Understand user intent |
| Slot Filling | Collect required information |
| Context Management | Maintain conversation context |
| Fulfillment | Integrate with AWS Lambda |
Lex Components
┌─────────────────────────────────────────────────────────────┐
│ Amazon Lex │
├─────────────────────────────────────────────────────────────┤
│ │
│ User Input ──▶ Intent Recognition ──▶ Slot Filling │
│ │ │ │
│ ▼ ▼ │
│ Dialog Management ──▶ Fulfillment (Lambda) │
│ │ │
│ ▼ │
│ Response │
│ │
└─────────────────────────────────────────────────────────────┘Lex Use Cases
| Use Case | Description |
|---|---|
| Customer Service Bots | Automated support |
| Order Taking | Food delivery, retail |
| Booking Systems | Hotel, flight reservations |
| Information Retrieval | FAQ bots |
| Productivity | Schedule management |
Lex Architecture
| Component | Description |
|---|---|
| Intent | What the user wants to do |
| Utterance | What user says/types |
| Slot | Data needed to fulfill intent |
| Fulfillment | Lambda function to complete action |
| Prompt | Question to get slot value |
Exam Tips
- Lex = Conversational AI for Chatbots
- Uses same technology as Alexa
- Supports voice and text
- ASR (speech recognition) + TTS (text-to-speech)
- Uses Lambda for backend logic
- Intents = user goals, Slots = parameters
9. Amazon Kendra
Overview
Amazon Kendra is an intelligent search service powered by machine learning.
Key Characteristics
| Feature | Description |
|---|---|
| Semantic Search | Understands meaning, not just keywords |
| Natural Language Queries | Ask questions in natural language |
| Multiple Data Sources | Search across repositories |
| Document Indexing | Automatic indexing and updates |
| Faceted Search | Filter by attributes |
| Answer Extraction | Extracts specific answers |
| Query Suggestions | Auto-complete suggestions |
Kendra Use Cases
| Use Case | Description |
|---|---|
| Enterprise Search | Search internal documents |
| Knowledge Base | FAQ, documentation search |
| Research | Legal, medical, financial research |
| Customer Support | Find answers for support agents |
| Compliance | Search policies, procedures |
Traditional Search vs Kendra
| Feature | Traditional Search | Amazon Kendra |
|---|---|---|
| Keyword Matching | ✅ | ✅ |
| Semantic Understanding | ❌ | ✅ |
| Natural Language | ❌ | ✅ |
| Answer Extraction | ❌ | ✅ |
| Learning | ❌ | ✅ |
Exam Tips
- Kendra = Intelligent Enterprise Search
- Uses ML for semantic search
- Understands natural language queries
- Goes beyond keyword matching
- Can extract specific answers from documents
10. Amazon Forecast
Overview
Amazon Forecast is a fully managed service for time-series forecasting.
Key Characteristics
| Feature | Description |
|---|---|
| Time-Series | Business metrics forecasting |
| ML-Based | Uses machine learning algorithms |
| Automatic | Auto-selects best algorithm |
| Explainability | Explains forecast drivers |
| Item-Level | Forecast per product/location |
Forecast Use Cases
| Use Case | Description |
|---|---|
| Demand Planning | Product demand forecasting |
| Inventory | Stock optimization |
| Resource Planning | Staffing, capacity planning |
| Financial | Revenue forecasting |
Exam Tips
- Forecast = Time-Series Forecasting Service
- Uses ML for predictions
- For business metrics analysis
- Used for demand, inventory, financial forecasting
11. Service Comparison and Selection
AI/ML Services Decision Tree
┌─────────────────────────────────────────────────────────────┐
│ Choose AI/ML Service │
├─────────────────────────────────────────────────────────────┤
│ │
│ What type of data? │
│ ├─ Images/Videos ──▶ Amazon Rekognition │
│ ├─ Text (Speech) ────▶ Amazon Transcribe (Speech→Text) │
│ │ Amazon Polly (Text→Speech) │
│ ├─ Text (Document) ─▶ Amazon Textract │
│ ├─ Text (Digital) ──▶ Amazon Comprehend (NLP) │
│ │ Amazon Lex (Chatbots) │
│ ├─ Documents Search ─▶ Amazon Kendra │
│ ├─ Time-Series ──────▶ Amazon Forecast │
│ └─ Custom ML Model ─▶ Amazon SageMaker AI │
│ │
└─────────────────────────────────────────────────────────────┘Service Quick Reference
| Task | Service |
|---|---|
| Identify objects in images | Amazon Rekognition |
| Detect faces/emotions | Amazon Rekognition |
| Convert speech to text | Amazon Transcribe |
| Convert text to speech | Amazon Polly |
| Translate languages | Amazon Translate |
| Extract text from documents | Amazon Textract |
| Analyze text sentiment | Amazon Comprehend |
| Build chatbot | Amazon Lex |
| Enterprise search | Amazon Kendra |
| Forecast metrics | Amazon Forecast |
| Build custom ML models | Amazon SageMaker AI |
12. Generative AI and Foundation Models (2026 Updates)
Amazon Bedrock
Amazon Bedrock is a fully managed service that makes foundation models (FMs) from leading AI companies available through an API.
Note: For CLF-C02, you only need to know Bedrock exists as a generative AI service. Details are for advanced certifications.
| Feature | Description |
|---|---|
| Foundation Models | Access to leading AI company FMs |
| Serverless | No infrastructure to manage |
| API-Based | Simple API integration |
| Security | Data privacy and security |
Bedrock Model Providers
| Provider | Models |
|---|---|
| AI21 Labs | Jurassic |
| Anthropic | Claude |
| Cohere | Command |
| Meta | Llama |
| Stability AI | Stable Diffusion |
| Amazon | Titan (Text, Embeddings) |
Generative AI Use Cases
| Use Case | Service |
|---|---|
| Text Generation | Bedrock (Titan, Claude) |
| Image Generation | Bedrock (Stable Diffusion) |
| Chatbots | Bedrock + Lex |
| Search | Kendra + Bedrock |
| Code Generation | Bedrock (Code Llama) |
Exam Tips - AI/ML Services
High-Yield Topics
Service Purposes:
- SageMaker AI = Build custom ML models
- Rekognition = Image/video analysis
- Transcribe = Speech to text
- Polly = Text to speech
- Translate = Language translation
- Textract = Document text/data extraction
- Comprehend = NLP text analysis
- Lex = Chatbots
- Kendra = Intelligent search
- Forecast = Time-series forecasting
Key Distinctions:
- Polly vs Transcribe = Text→Speech vs Speech→Text
- Textract vs Comprehend = Document extraction vs Text analysis
- Rekognition = Images/Videos only
- SageMaker AI vs others = Custom models vs purpose-built services
Integrations:
- S3 = Storage for Rekognition, Textract
- SNS = Async notifications (Rekognition)
- Lambda = Backend logic (Lex)
2026 Update:
- Bedrock = Generative AI foundation models
- Titan, Claude, Llama, Stable Diffusion
Additional Resources
DigitalCloud Training Cheat Sheets
- AWS Machine Learning Services Cheat Sheet - Comprehensive AI/ML services reference for exam prep
Official AWS Documentation
- Amazon SageMaker AI
- Amazon Rekognition
- Amazon Transcribe
- Amazon Polly
- Amazon Translate
- Amazon Textract
- Amazon Comprehend
- Amazon Lex
- Amazon Kendra
- Amazon Forecast
- Amazon Bedrock (Generative AI)
Practice Resources
Previous: Analytics Services