What Are AI Use Cases?
AI is already part of everyday life — YouTube recommends your next video, spam emails get sorted automatically, and your phone camera recognizes faces. All of these are AI use cases.
The AWS AIF-C01 exam tests whether you can identify which type of problem AI can solve and which AWS service fits best. This post organizes AI use cases by type and maps them to the right AWS services.
---
Types of Problems AI Can Solve
AI excels in five main categories.
Classification — "What is this?"
Sorting input data into predefined categories. Examples: spam detection, image labeling, sentiment analysis.
| AWS Service | Use Case | |---|---| | Amazon Comprehend | Text sentiment, language detection | | Amazon Rekognition | Image/video object classification | | Amazon SageMaker | Custom classification model training |
Prediction — "What will happen next?"
Forecasting future values from historical data. Examples: sales forecasting, churn prediction.
| AWS Service | Use Case | |---|---| | Amazon Forecast | Time-series demand forecasting | | Amazon Personalize | Personalized recommendations | | Amazon SageMaker | Custom prediction models |
Generation — "Create something new"
Producing new text, images, code, or audio. Examples: chatbots, product description writing, code completion.
| AWS Service | Use Case | |---|---| | Amazon Bedrock | LLM-based text and image generation | | Amazon Q | Enterprise AI assistant | | Amazon Titan | AWS-native foundation models |
Detection — "Find anomalies"
Spotting patterns that deviate from normal. Examples: fraud detection, equipment failure prediction.
| AWS Service | Use Case | |---|---| | Amazon Fraud Detector | Online fraud detection | | Amazon Lookout for Metrics | Business metric anomaly detection | | Amazon Lookout for Equipment | Industrial equipment anomaly detection |
Natural Language Processing — "Understand human language"
Processing and understanding text and speech.
| AWS Service | Use Case | |---|---| | Amazon Comprehend | Entity recognition, sentiment analysis | | Amazon Transcribe | Speech-to-text | | Amazon Translate | Automatic translation | | Amazon Lex | Conversational chatbots | | Amazon Polly | Text-to-speech |
!5 types of AI problems and the AWS services that solve them
Keyword-to-Service Mapping
| Scenario Keyword | AWS Service | |---|---| | Face/object recognition in images | Amazon Rekognition | | Text sentiment analysis | Amazon Comprehend | | Speech to text | Amazon Transcribe | | Text to speech | Amazon Polly | | Automatic translation | Amazon Translate | | Build a chatbot | Amazon Lex | | Demand/inventory forecasting | Amazon Forecast | | Personalized recommendations | Amazon Personalize | | Fraud detection | Amazon Fraud Detector | | Large language models | Amazon Bedrock | | Custom ML model pipeline | Amazon SageMaker | | Extract data from documents | Amazon Textract | | Medical text analysis | Amazon Comprehend Medical |
---
Exam Key Takeaways
Image/video analysis → Rekognition Text analysis/sentiment → Comprehend Speech recognition → Transcribe, Speech synthesis → Polly Translation → Translate Chatbots → Lex Time-series forecasting → Forecast Recommendations → Personalize Fraud/anomaly detection → Fraud Detector / Lookout series Document data extraction → Textract Generative AI / LLMs → Bedrock Custom ML pipeline → SageMaker Enterprise AI assistant → Amazon Q
On the exam, "most suitable service" questions are all about spotting keywords. Memorize the table above and you can handle most questions.