Claude Certified Associate — Foundations (CCAO-F): complete study guide
The CCAO-F is Anthropic's entry-level certification and the only one of the four written for people who do not write code. It certifies that you can use Claude competently and responsibly as a working professional — which, as the domain weights show, is mostly a test of judgement rather than of prompting.
What this exam is actually for
The CCAO-F is the entry point to Anthropic's certification programme and the only one of the four written for people who do not code. It is aimed at operations, marketing, project management, communications, education and general knowledge work — people who use Claude as a serious daily tool and need to prove they use it competently and safely.
If you write software for a living, this is not your exam: the Developer and Architect tracks cover that ground. What CCAO-F tests is judgement. Can you write a prompt that gets the job done, tell whether the answer is good enough to send, pick the right Claude feature for the task, fit it into a real business process, and stay on the right side of your organisation's policies while doing it?
Format
| Questions | 60 | Time | 120 minutes |
|---|---|---|---|
| Pass mark | 720 / 1000 (scaled) | Format | Multiple choice, one correct answer |
| Domains | 7 | Credential validity | 12 months |
Two hours for sixty questions is two minutes each, which is generous — the pressure on this exam is not time. It is that the answer options are all things a reasonable person might do. Scoring is scaled rather than a raw percentage, so treat 80% on a realistic practice set as your comfort threshold rather than aiming at 72%.
Domain weights
| # | Domain | Weight | Roughly | The question behind it |
|---|---|---|---|---|
| 2 | Output Evaluation and Validation | 21% | ~13 questions | Is this good enough to send? |
| 4 | Workflow Integration and Solution Design | 16% | ~10 questions | Where does this fit in how we already work? |
| 6 | Governance, Risk, and Responsible Use | 15% | ~9 questions | Am I allowed to do this with this information? |
| 1 | Prompting and Task Execution | 14% | ~8 questions | How do I ask for what I actually want? |
| 3 | Product and Model Selection | 12% | ~7 questions | Which Claude feature and which model? |
| 5 | Configuration and Knowledge Management | 12% | ~7 questions | How do I set this up once so it keeps working? |
| 7 | Troubleshooting and Optimisation | 10% | ~6 questions | It gave me something poor. Now what? |
Note what the weighting is telling you. Prompting — the thing most people assume an AI certification is about — is only 14%. Evaluating the output is half again as heavy, at 21%. Governance is worth more than prompting too. This is a exam about being a responsible operator, not a clever prompter.
How the questions are written
Almost every question is a short workplace scenario: a marketing manager preparing a campaign brief, an HR coordinator screening a policy document, an analyst summarising customer feedback. You are given a situation and four courses of action, and asked which is best. The wrong options are rarely absurd. They are usually one of these three shapes:
- Faster but unverified — it gets the task done and skips the check that the situation required.
- Thorough but disproportionate — it applies heavyweight process to a low-stakes task, which in a real workplace means nobody uses it.
- Technically fine but out of bounds — it would work, but it puts confidential data somewhere it should not go, or makes a decision a human must own.
Find the stakes in the stem. The correct answer is almost always the one whose level of caution matches the consequence of being wrong.
Every page in this guide
- The mental model — One principle sits under most CCAO-F questions: Claude produces the draft, but you remain accountable for anything you send, file or decide.
- Domain 1 — Prompting and Task Execution — Writing prompts that work: task, audience, format and constraints, examples, decomposing large requests, and iterating instead of regenerating.
- Domain 2 — Output Evaluation and Validation — The heaviest CCAO-F domain: accuracy, completeness and appropriateness, spotting hallucinations, six techniques for checkable output, and matching verification to stakes.
- Domain 3 — Product and Model Selection — Choosing between Projects, Research, Artifacts, connectors and uploads, picking a model tier proportionate to the task, and working within context limits.
- Domain 4 — Workflow Integration and Solution Design — Finding the right task for Claude, designing a workflow around a review step, integrating rather than replacing, and communicating value and limitations honestly.
- Domain 5 — Configuration and Knowledge Management — Setting up Projects that keep working: writing instructions, curating knowledge, choosing connectors over uploads, and maintaining a shared configuration.
- Domain 6 — Governance, Risk and Responsible Use — Data handling on consumer versus commercial plans, prohibited uses, the eight high-risk domains requiring human review and disclosure, and confidentiality traps.
- Domain 7 — Troubleshooting and Optimisation — Diagnosing a poor result from its symptom, working efficiently, and turning individual fixes into team-wide improvement.
- The 12 anti-patterns — Twelve distractors that recur across the CCAO-F question bank, each with why it fails and what the exam wants instead.
- Cheat sheet — Domain weights, which feature for which job, model tiers, verification levels, the eight high-risk domains and the data-handling rules.
- Questions & answers — Domain 1 — CCAO-F practice questions on prompting and task execution, each with the correct answer and a full explanation of why the distractors fail.
- Questions & answers — Domain 2 — CCAO-F practice questions on output evaluation and validation, each with the correct answer and a full explanation of why the distractors fail.
- Questions & answers — Domain 3 — CCAO-F practice questions on product and model selection, each with the correct answer and a full explanation of why the distractors fail.
- Questions & answers — Domain 4 — CCAO-F practice questions on workflow integration and solution design, each with the correct answer and a full explanation of why the distractors fail.
- Questions & answers — Domain 5 — CCAO-F practice questions on configuration and knowledge management, each with the correct answer and a full explanation of why the distractors fail.
- Questions & answers — Domain 6 — CCAO-F practice questions on governance, risk and responsible use, each with the correct answer and a full explanation of why the distractors fail.
- Questions & answers — Domain 7 — CCAO-F practice questions on troubleshooting and optimisation, each with the correct answer and a full explanation of why the distractors fail.
- Facts to memorise — Thirty recall facts covering the evaluation dimensions, high-risk domains, feature choices, model tiers and data-handling defaults.
- Four-week study plan — A week-by-week CCAO-F preparation plan built around practising judgement rather than memorising features.
- Official resources — Anthropic Academy, the Claude Help Center, the Usage Policy, the Privacy Center and Pearson VUE, mapped to the seven exam domains.
This is an independent study guide, not an official Anthropic publication. Exam blueprints and product details change; confirm the current domain weights and format against the official exam guide before you sit.
Next: The mental model →
Explore Other Certification Study Guides
Switch to another exam guide to prepare for your next Claude certification:
Claude Certified Architect — Foundations
Core architectural principles: agentic loops & orchestration, tool design & MCP, Claude Code workflows, prompt engineering, and context reliability.
Claude Certified Architect — Professional
Advanced enterprise architecture: solution design, system integration, evaluation frameworks, governance, safety & developer enablement.
Claude Certified Developer — Foundations
Build software on Claude: API integration, model selection, tools & MCP servers, agents vs workflows, security, safety & debugging.