The AWS Certified AI Practitioner (AIF-C01) is Amazon Web Services' foundational certification for anyone who works with—or alongside—AI and machine learning solutions on AWS. The exam is organized into five official domains, each carrying a specific percentage of the total score, and knowing those weightings is the single fastest way to prioritize your study time. In this guide we break down every domain, explain what it actually tests, and show you where to invest the most effort.
Why Domain Weightings Matter for Your Study Plan
Before diving into the content, it's worth understanding why the domain breakdown is so valuable. The AIF-C01 exam draws questions proportionally from each domain. A domain worth 28% of the exam will contribute roughly twice as many questions as one worth 14%. If you spend equal time on every topic, you're leaving easy points on the table.
Here's a quick reference table of all five domains:
| # | Domain | Weighting |
|---|---|---|
| 1 | Fundamentals of AI and ML | 20% |
| 2 | Fundamentals of Generative AI | 24% |
| 3 | Applications of Foundation Models | 28% |
| 4 | Guidelines for Responsible AI | 14% |
| 5 | Security, Compliance, and Governance for AI Solutions | 14% |
Domains 2 and 3 together account for 52% of the exam—more than half. That doesn't mean you can ignore the others, but it does mean generative AI and foundation models deserve the lion's share of your preparation.
Domain 1: Fundamentals of AI and ML — 20%
This domain is the foundation everything else is built on. Before AWS can test you on cutting-edge generative AI, it needs to confirm you understand the core concepts that underpin all machine learning systems.
What This Domain Covers
Core AI and ML concepts — You should be able to explain the difference between artificial intelligence, machine learning, and deep learning. Think of these as nested circles: AI is the broadest concept, ML is a subset that learns from data, and deep learning is a subset of ML that uses neural networks with many layers.
Types of ML problems — The exam expects you to distinguish between supervised learning (labeled training data), unsupervised learning (finding patterns in unlabeled data), and reinforcement learning (learning through reward signals). You should also recognize common use cases for each, such as classification, regression, clustering, and recommendation systems.
The ML development lifecycle — From data collection and preprocessing through model training, evaluation, and deployment, you need a high-level understanding of how a model goes from idea to production. AWS services like Amazon SageMaker sit at the center of this lifecycle on the AWS platform.
Key ML terminology — Terms like features, labels, training sets, validation sets, test sets, overfitting, underfitting, and hyperparameters will appear throughout the exam. Domain 1 is where you build the vocabulary you'll need to understand every other domain.
Study Tips for Domain 1
- If you're new to ML, spend time here first—it unlocks the rest of the material.
- Focus on conceptual understanding rather than math. This is a practitioner-level exam, not a data science exam.
- Map AWS services (SageMaker, Rekognition, Comprehend, Forecast) to the ML problem types they solve.
Domain 2: Fundamentals of Generative AI — 24%
Generative AI is the reason this certification exists. AWS created the AIF-C01 specifically to validate knowledge of AI technologies that generate text, images, code, and other content—and this domain lays the conceptual groundwork for that conversation.
What This Domain Covers
What generative AI is and how it differs from traditional ML — Traditional ML models are trained to predict or classify. Generative models are trained to create. Understanding this distinction—and why it matters for business use cases—is central to Domain 2.
Large language models (LLMs) — You need to understand what an LLM is, how it's trained on massive text corpora, and why scale matters. Concepts like tokens, context windows, and temperature (which controls output randomness) are fair game.
Transformer architecture at a conceptual level — You won't be asked to implement a transformer, but you should understand that the attention mechanism is what allows modern LLMs to understand relationships between words across long passages of text.
Common generative AI use cases — Text summarization, question answering, code generation, image synthesis, and conversational AI (chatbots) are all examples the exam may reference. Being able to match a business scenario to the right generative AI capability is a key skill.
AWS generative AI services — Amazon Bedrock is the flagship service here. Know that Bedrock provides access to foundation models from multiple providers through a single API, without requiring you to manage the underlying infrastructure.
Study Tips for Domain 2
- Spend real time with Amazon Bedrock's documentation and the AWS Generative AI learning path.
- Practice explaining generative AI concepts in plain language—the exam often presents business scenarios and asks you to identify the right approach.
- Understand the difference between training a model from scratch, fine-tuning an existing model, and using a model via API with prompt engineering.
Domain 3: Applications of Foundation Models — 28%
This is the highest-weighted domain on the AIF-C01 exam, and for good reason. Foundation models—large, pre-trained models that can be adapted to a wide range of tasks—are the practical engine behind most real-world generative AI applications. Domain 3 tests your ability to apply these models effectively.
What This Domain Covers
Prompt engineering — How you phrase a question or instruction to an LLM dramatically affects the quality of the output. The exam covers prompt engineering techniques including zero-shot prompting (no examples provided), few-shot prompting (a handful of examples in the prompt), and chain-of-thought prompting (asking the model to reason step by step). Understanding when to use each technique is essential.
Retrieval-Augmented Generation (RAG) — RAG is one of the most important patterns in enterprise generative AI. Instead of relying solely on what a model learned during training, RAG retrieves relevant documents from an external knowledge base and includes them in the prompt context. This allows the model to answer questions about proprietary or up-to-date information it was never trained on. AWS supports RAG through Amazon Bedrock Knowledge Bases.
Fine-tuning and customization — Sometimes prompt engineering isn't enough. Fine-tuning involves continuing to train a foundation model on a smaller, domain-specific dataset so it learns your organization's terminology, tone, or specialized knowledge. The exam distinguishes between fine-tuning (updating model weights) and RAG (augmenting the prompt at inference time).
Agents and multi-step reasoning — Amazon Bedrock Agents allow foundation models to take actions—calling APIs, querying databases, running code—in order to complete complex, multi-step tasks. Understanding the concept of an AI agent and how it differs from a simple chatbot is increasingly important.
Evaluating foundation model outputs — How do you know if your model is performing well? Domain 3 covers evaluation approaches including human evaluation, automated metrics, and the challenge of hallucination—when a model confidently produces factually incorrect information.
Choosing the right foundation model — Not every task requires the most powerful (and expensive) model. The exam tests your ability to consider factors like latency, cost, context window size, and task complexity when selecting a model for a given use case.
Study Tips for Domain 3
- This domain rewards hands-on practice. Use the Amazon Bedrock console to experiment with different models and prompting strategies.
- Build a mental model of the RAG pipeline: user query → vector search → retrieved context → augmented prompt → model response.
- Understand the trade-offs between RAG and fine-tuning. RAG is faster and cheaper to implement; fine-tuning produces a more deeply customized model but requires more data and compute.
- Review the Amazon Bedrock Agents documentation to understand how agents orchestrate multi-step workflows.
Domain 4: Guidelines for Responsible AI — 14%
As AI systems become more powerful, the potential for harm grows alongside the potential for benefit. Domain 4 tests your understanding of the principles and practices that help organizations deploy AI responsibly.
What This Domain Covers
Fairness and bias — ML models learn from historical data, and if that data reflects societal biases, the model will too. The exam covers how bias can enter a model (through biased training data, biased labels, or biased feature selection) and strategies for detecting and mitigating it.
Explainability and transparency — Stakeholders—whether customers, regulators, or internal teams—often need to understand why a model made a particular decision. Tools like Amazon SageMaker Clarify help provide explanations for model predictions. The exam distinguishes between interpretable models (like decision trees) and black-box models (like deep neural networks) where explainability requires additional tooling.
Robustness and reliability — A responsible AI system should perform consistently across different populations and edge cases. Domain 4 covers concepts like model drift (when a model's performance degrades over time as real-world data changes) and the importance of ongoing monitoring.
Human oversight — Responsible AI doesn't mean fully autonomous AI. The exam emphasizes the importance of keeping humans in the loop, especially for high-stakes decisions in areas like healthcare, finance, and hiring.
AWS responsible AI principles — AWS has published its own responsible AI framework. Familiarize yourself with AWS's stated principles around fairness, explainability, privacy, security, and robustness.
Study Tips for Domain 4
- Read AWS's responsible AI documentation and any published whitepapers on the topic.
- Think through real-world scenarios: What could go wrong if a hiring algorithm is biased? What happens when a medical diagnosis model drifts?
- Understand the specific AWS tools designed to support responsible AI, particularly SageMaker Clarify and SageMaker Model Monitor.
Domain 5: Security, Compliance, and Governance for AI Solutions — 14%
The final domain addresses the organizational and technical controls needed to deploy AI safely and in compliance with regulations. This domain overlaps with general AWS security knowledge but applies it specifically to AI workloads.
What This Domain Covers
Data security for AI — Training data and model outputs can contain sensitive information. The exam covers encryption at rest and in transit, data classification, and the importance of not exposing personally identifiable information (PII) in prompts sent to third-party models.
Access control and identity — AWS Identity and Access Management (IAM) applies to AI services just as it does to any other AWS resource. You should understand how to use IAM roles and policies to control who can invoke a model, access a knowledge base, or modify a fine-tuned model.
Model governance — Organizations need processes for tracking which models are in production, what data they were trained on, and how they're performing. This is sometimes called an AI model registry or model catalog. Amazon SageMaker Model Registry supports this use case.
Compliance and regulatory considerations — Depending on the industry, AI systems may be subject to regulations around data privacy (like GDPR or HIPAA), algorithmic accountability, or sector-specific rules. The exam tests your awareness of these considerations and how AWS services can help meet compliance requirements.
Threat modeling for AI — AI systems face unique security threats, including prompt injection (where malicious input manipulates a model's behavior), data poisoning (corrupting training data to influence model outputs), and model inversion attacks (attempting to extract training data from a model). Understanding these threats at a conceptual level is part of Domain 5.
Study Tips for Domain 5
- If you have an AWS security background, this domain will feel familiar—apply what you know about IAM, encryption, and VPCs to AI-specific scenarios.
- Pay special attention to AI-specific threats like prompt injection, which are unique to generative AI systems.
- Review AWS compliance programs (SOC 2, HIPAA eligibility, etc.) and understand how they apply to services like Amazon Bedrock and SageMaker.
How to Allocate Your Study Time
Based on the domain weightings, here's a suggested time allocation for a typical 40-hour study plan:
| Domain | Weighting | Suggested Study Hours |
|---|---|---|
| 1. Fundamentals of AI and ML | 20% | 8 hours |
| 2. Fundamentals of Generative AI | 24% | 10 hours |
| 3. Applications of Foundation Models | 28% | 11 hours |
| 4. Guidelines for Responsible AI | 14% | 5.5 hours |
| 5. Security, Compliance, and Governance | 14% | 5.5 hours |
These are starting points, not rigid rules. If you already have a strong security background, you can shift time from Domain 5 toward Domain 3. If you're brand new to AI concepts, invest more heavily in Domain 1 before moving on.
Putting It All Together
The AWS Certified AI Practitioner (AIF-C01) is designed for a broad audience—business analysts, project managers, developers, and anyone who needs to understand AI and ML in the context of AWS—not just data scientists. The five domains reflect that breadth, moving from foundational concepts through practical application and ending with the governance and security considerations that matter in real enterprise deployments.
The clearest path to passing is to:
- Build your conceptual foundation in Domain 1 before tackling the generative AI content.
- Invest the most time in Domains 2 and 3, which together make up more than half the exam.
- Don't neglect Domains 4 and 5—responsible AI and governance questions are increasingly common in real-world AI projects, and 28% of the exam lives there.
- Practice with realistic questions that mirror the scenario-based format AWS uses.
Ready to Test Your Knowledge?
Reading about exam domains is a great start, but the fastest way to find your weak spots is to answer practice questions under realistic conditions. LearnZapp offers a free AWS Certified AI Practitioner practice test built around the official AIF-C01 domain structure—so every question you answer maps directly to the domains and weightings covered in this guide. Try it free today and find out exactly where to focus your final study sessions before exam day.