AI service
Amazon Comprehend
Natural language processing for sentiment, entities, key phrases, language, PII, and topics.
Key points
- Provides prebuilt natural language processing APIs for text.
- Can detect sentiment, entities, key phrases, language, syntax, topics, and personally identifiable information.
- Supports custom classification and custom entity recognition for domain-specific text needs.
- Works on text input rather than images, audio, or document layout.
- Reduces the need to train a custom NLP model for common text analysis tasks.
When to use it
- Choose Comprehend to analyze customer reviews, support tickets, emails, or social text.
- Use it for sentiment analysis, PII detection, entity extraction, or topic discovery.
- Use custom Comprehend features when standard NLP labels are not enough.
Exam tips
- Comprehend analyzes text meaning; Textract extracts text and structure from documents.
- Kendra searches enterprise content; Comprehend classifies and extracts information from text.
- Translate changes language, while Comprehend detects language and analyzes the text.
- For medical text, AWS has specialized health AI services, but AIF-C01 usually tests the general service map.