Skip to content

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

CategoryServicesPurpose
ML PlatformsSageMaker AIBuild, train, deploy custom ML models
Computer VisionRekognitionImage and video analysis
NLP - Text AnalysisComprehendExtract insights from text
NLP - SpeechTranscribe, Polly, TranslateSpeech-to-text, text-to-speech, translation
Document AnalysisTextractExtract data from documents
Conversational AILexBuild chatbots
SearchKendraIntelligent enterprise search
ForecastingForecastTime-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

CapabilityDescription
Data LabelingGround truth labeling for training data
Feature EngineeringPrepare, transform data
Bias DetectionDetect statistical bias in data and models
AutoMLAutomatically build models
TrainingManaged training infrastructure
TuningHyperparameter optimization
HostingDeploy models with auto scaling
MonitoringModel performance monitoring
WorkflowsML 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 CaseSageMaker Feature
Custom ML model developmentStudio Notebooks
Automated model buildingSageMaker Autopilot
Large-scale trainingManaged training jobs
Real-time predictionsSageMaker Endpoints
Batch predictionsBatch Transform
MLOps automationSageMaker Pipelines

SageMaker AI vs Other AI Services

ScenarioUse SageMaker AIUse Purpose-Built Service
Custom ML model✅ Yes❌ No
Image analysisPossible, but use Rekognition✅ Amazon Rekognition
Text-to-speechPossible, but use Polly✅ Amazon Polly
ChatbotPossible, 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

CapabilityDescription
Object DetectionIdentify objects (cars, pets, furniture)
Scene RecognitionIdentify scenes (beach, mountain, city)
Face AnalysisDetect faces, emotions, age range
Face ComparisonVerify if two faces match
Text DetectionExtract text from images (OCR)
Celebrity RecognitionIdentify celebrities in images
Unsafe ContentDetect inappropriate content
Custom LabelsTrain custom models for your objects

Rekognition Use Cases

Use CaseDescription
Content ModerationDetect inappropriate images/videos
Face VerificationUser authentication (ID verification)
Sentiment AnalysisDetect emotions from faces
Document ProcessingExtract text from documents
Asset ManagementSearch images by content
SecurityDetect 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

FeatureDescription
Speech-to-TextConvert audio to text
Deep LearningAdvanced ASR technology
Multiple LanguagesSupport for global languages
Real-TimeLive transcription
BatchProcess recorded audio files
Speaker IdentificationDistinguish between speakers
Automatic PunctuationAdds punctuation automatically
Custom VocabularyAdd domain-specific terms

Transcribe Use Cases

Use CaseDescription
Call TranscriptionCustomer support calls
Meeting NotesAutomatic meeting transcription
CaptioningVideo subtitles
Voice AnalyticsAnalyze customer conversations
DocumentationMedical/legal dictation

Input/Output Formats

InputOutput
Audio files (WAV, MP3, MP4)Text transcripts
Live audio streamsReal-time text
Phone recordingsTimestamped 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

FeatureDescription
Text-to-SpeechConvert text to audio
Lifelike VoicesNatural-sounding speech
Multiple LanguagesSupport for global languages
SSML SupportSpeech Synthesis Markup Language
Neural TTSAdvanced deep learning voices
Voice CustomizationAdjust pitch, rate, volume
Newscaster StyleNews reading style

Polly Use Cases

Use CaseDescription
AccessibilityVisual impairment assistance
Voice AssistantsAlexa-style applications
E-learningCourse narration
AudiobooksText-to-audio conversion
GamingCharacter voices
IVR SystemsInteractive voice response

Polly vs Transcribe

ServiceDirectionUse Case
PollyText → SpeechGive your app a voice
TranscribeSpeech → TextConvert 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

FeatureDescription
Neural MLDeep learning models
Language Pairs75+ language pairs
Batch TranslationTranslate large volumes
Real-TimeLive translation
Custom TerminologyDomain-specific vocabulary
FormalityFormal/informal tones

Translate Use Cases

Use CaseDescription
Website LocalizationMulti-language websites
Document TranslationTranslate documents
Customer SupportMulti-language support
Content DistributionGlobal content reach
CommunicationCross-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

CapabilityDescription
OCROptical Character Recognition
HandwritingRecognizes printed and handwritten text
Forms DataExtract key-value pairs from forms
Table DataExtract table structures
Document AnalysisUnderstands document context
Identity DocumentsExtract from IDs, passports
Invoices/ReceiptsUnderstand business documents
RelationshipsIdentifies relationships in data

Textract vs OCR

FeatureTraditional OCRAmazon Textract
Text Only
Handwriting
Form Fields
Tables
Context Understanding
Document Type Awareness

Textract Use Cases

Use CaseDescription
Invoice ProcessingExtract line items, totals
Form AutomationProcess application forms
Document SearchSearch within scanned PDFs
ComplianceKYC, identity verification
Data EntryAutomate manual data entry
Receipt ProcessingExpense management

Supported Document Formats

FormatExamples
ImagesPNG, JPG, TIFF
PDFsScanned and native PDF
FormsTax forms, applications
TablesFinancial 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

CapabilityDescription
Sentiment AnalysisDetect positive, negative, neutral, mixed
Key Phrase ExtractionFind important phrases
Entity RecognitionIdentify people, places, dates, quantities
Topic ModelingDiscover topics in document collections
Language DetectionIdentify document language
PII DetectionDetect personally identifiable information
Syntax AnalysisParse sentence structure

Comprehend Use Cases

Use CaseDescription
Customer Feedback AnalysisUnderstand customer sentiment
Document CategorizationAuto-categorize documents
Social Media MonitoringTrack brand sentiment
ComplianceDetect PII in documents
Knowledge DiscoveryFind topics in large document sets
Content ModerationDetect inappropriate content

Comprehend vs Textract

ServiceInputOutput
ComprehendText (already digital)Insights (sentiment, entities, topics)
TextractScanned documents/imagesExtracted 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

CapabilityDescription
ChatbotsBuild conversational bots
Voice & TextBoth modalities supported
ASR & TTSSpeech recognition and synthesis
Intent RecognitionUnderstand user intent
Slot FillingCollect required information
Context ManagementMaintain conversation context
FulfillmentIntegrate with AWS Lambda

Lex Components

┌─────────────────────────────────────────────────────────────┐
│                       Amazon Lex                             │
├─────────────────────────────────────────────────────────────┤
│                                                              │
│  User Input ──▶ Intent Recognition ──▶ Slot Filling          │
│                      │                    │                   │
│                      ▼                    ▼                   │
│              Dialog Management ──▶ Fulfillment (Lambda)       │
│                      │                                       │
│                      ▼                                       │
│                  Response                                   │
│                                                              │
└─────────────────────────────────────────────────────────────┘

Lex Use Cases

Use CaseDescription
Customer Service BotsAutomated support
Order TakingFood delivery, retail
Booking SystemsHotel, flight reservations
Information RetrievalFAQ bots
ProductivitySchedule management

Lex Architecture

ComponentDescription
IntentWhat the user wants to do
UtteranceWhat user says/types
SlotData needed to fulfill intent
FulfillmentLambda function to complete action
PromptQuestion 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

FeatureDescription
Semantic SearchUnderstands meaning, not just keywords
Natural Language QueriesAsk questions in natural language
Multiple Data SourcesSearch across repositories
Document IndexingAutomatic indexing and updates
Faceted SearchFilter by attributes
Answer ExtractionExtracts specific answers
Query SuggestionsAuto-complete suggestions

Kendra Use Cases

Use CaseDescription
Enterprise SearchSearch internal documents
Knowledge BaseFAQ, documentation search
ResearchLegal, medical, financial research
Customer SupportFind answers for support agents
ComplianceSearch policies, procedures

Traditional Search vs Kendra

FeatureTraditional SearchAmazon 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

FeatureDescription
Time-SeriesBusiness metrics forecasting
ML-BasedUses machine learning algorithms
AutomaticAuto-selects best algorithm
ExplainabilityExplains forecast drivers
Item-LevelForecast per product/location

Forecast Use Cases

Use CaseDescription
Demand PlanningProduct demand forecasting
InventoryStock optimization
Resource PlanningStaffing, capacity planning
FinancialRevenue 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

TaskService
Identify objects in imagesAmazon Rekognition
Detect faces/emotionsAmazon Rekognition
Convert speech to textAmazon Transcribe
Convert text to speechAmazon Polly
Translate languagesAmazon Translate
Extract text from documentsAmazon Textract
Analyze text sentimentAmazon Comprehend
Build chatbotAmazon Lex
Enterprise searchAmazon Kendra
Forecast metricsAmazon Forecast
Build custom ML modelsAmazon 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.

FeatureDescription
Foundation ModelsAccess to leading AI company FMs
ServerlessNo infrastructure to manage
API-BasedSimple API integration
SecurityData privacy and security

Bedrock Model Providers

ProviderModels
AI21 LabsJurassic
AnthropicClaude
CohereCommand
MetaLlama
Stability AIStable Diffusion
AmazonTitan (Text, Embeddings)

Generative AI Use Cases

Use CaseService
Text GenerationBedrock (Titan, Claude)
Image GenerationBedrock (Stable Diffusion)
ChatbotsBedrock + Lex
SearchKendra + Bedrock
Code GenerationBedrock (Code Llama)

Exam Tips - AI/ML Services

High-Yield Topics

  1. 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
  2. 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
  3. Integrations:

    • S3 = Storage for Rekognition, Textract
    • SNS = Async notifications (Rekognition)
    • Lambda = Backend logic (Lex)
  4. 2026 Update:

    • Bedrock = Generative AI foundation models
    • Titan, Claude, Llama, Stable Diffusion

Additional Resources

DigitalCloud Training Cheat Sheets

Official AWS Documentation

Practice Resources


Previous: Analytics Services

Released under the MIT License.