AIF-C01
Deep Dive
AWS services that add AI capabilities to your applications through a simple API call, without needing to train or manage any model. The exam constantly tests whether you can match the correct scenario to the correct service.
Contents
Analyzes images and video with Machine Learning: object, scene, and activity detection, face recognition and comparison, content moderation, and text detection within an image.
Concrete use case
An identity-verification app that compares an ID photo with a live selfie in real time to confirm it's the same person.
What tends to be confused on the exam
Rekognition analyzes the VISUAL CONTENT of an image (what's in it). It does not extract structured text from documents like forms or tables — that's Textract's job, even though both can "read" text in an image.
Detect objects and scenes in an image using Amazon Rekognition, without training your own vision model.
+1 more...
Extracts text, tables, and form data from scanned documents or PDFs, understanding the document's structure — not just plain text like a traditional OCR.
Concrete use case
An insurer that automatically processes thousands of scanned claim forms, extracting each field (name, date, amount) as structured data.
What tends to be confused on the exam
Textract extracts and structures text from documents. It doesn't understand the MEANING of the text (sentiment, entities, topics) — for that, once the text is extracted, you'd combine it with Comprehend.
Extract key fields from a scanned invoice using Amazon Textract, without templates or manual OCR.
+1 more...
Automatic speech recognition (ASR): converts audio into text, with support for speaker identification, custom vocabulary, and multiple languages.
Concrete use case
A call center that automatically transcribes every call to later analyze it and detect recurring complaints.
What tends to be confused on the exam
Transcribe converts AUDIO into TEXT. It doesn't analyze that text's content (that would be Comprehend) nor does it do the reverse process, text to speech (that's Polly).
Transcribe a support call to text with speaker separation using Amazon Transcribe.
+1 more...
Converts text into realistic speech (Text-to-Speech). Offers natural-sounding voices in multiple languages and dialects, including more expressive neural voices.
Concrete use case
A language-learning app that reads lesson content aloud with native pronunciation.
What tends to be confused on the exam
Polly is the REVERSE process of Transcribe: text → speech, not speech → text. It's easy to mix up the two on the exam because their descriptions sound similar.
Generate audio from text with a neural voice from Amazon Polly.
+1 more...
Extracts information and relationships from unstructured text: sentiment analysis (positive/negative/neutral/mixed), entity extraction (people, places, dates, quantities), language detection, and topic classification.
Concrete use case
Automatically analyzing thousands of customer reviews to identify overall sentiment and the most-mentioned topics without reading them one by one.
What tends to be confused on the exam
Comprehend analyzes the MEANING of existing text. It doesn't transcribe audio (Transcribe), doesn't extract data from scanned documents (Textract), and doesn't generate conversation (Lex).
Analyze sentiment and extract entities from a text with Amazon Comprehend, without training any model.
+1 more...
Builds conversational interfaces (chatbots and voice assistants) using the same technology as Amazon Alexa. Defines intents and slots (parameters), and manages multi-turn conversations.
Concrete use case
A chatbot on the website that answers frequently asked questions about orders and escalates to a human agent when it can't resolve the query.
What tends to be confused on the exam
Lex is for building STRUCTURED conversational flows with predefined intents. It's not the same as a Bedrock agent, which reasons openly with a foundation model and can dynamically orchestrate multiple tools.
Create a conversational bot in Amazon Lex and test it by recognizing an intent and its slots.
+1 more...
Enterprise search engine powered by Machine Learning. Searches across documents, FAQs, and multiple data repositories, returning direct, relevant answers, not just a list of links like a traditional search engine.
Concrete use case
A company with thousands of internal documents where employees need to quickly find specific information without knowing which document holds it.
What tends to be confused on the exam
Kendra is ready-to-use semantic search over your repositories. It's not the same as building your own RAG system with Bedrock Knowledge Bases, even though both solve the "search your own documents" problem — Kendra is a search service, Knowledge Bases is designed to feed a generative foundation model.
Create a Kendra index and run a natural-language search over company documents.
+1 more...
| Exam scenario | Correct service |
|---|---|
| Detect objects, faces, or inappropriate content in images/video | Amazon Rekognition |
| Extract fields and tables from scanned forms or PDFs | Amazon Textract |
| Transcribe calls or audio recordings to text | Amazon Transcribe |
| Read content aloud / generate audio from text | Amazon Polly |
| Analyze sentiment or extract entities from reviews and comments | Amazon Comprehend |
| Build a chatbot with predefined intents and slots | Amazon Lex |
| Search for answers within the company's internal documents | Amazon Kendra |
No guided steps: read a scenario and pick, among 4 AWS AI services, the only one that meets every requirement.
+1 more...
Did you get this topic?
Test what you just learned
A logistics company receives hundreds of scanned invoices per day in different formats and needs to automatically extract the invoice number, date, and total amount as structured data. Which AWS service should they use?
Put what you learned to the test with original AIF-C01 exam questions.
Start AIF-C01 mock examUnlimited mock exams with the Pro plan
Log in to track your progress.