Claude Certified Associate — Foundations
The non-technical Claude certification: knowing when an answer cannot be trusted, validating it properly, wiring Claude into real work, and staying inside the Usage Policy — as scenario flashcards you actually remember.
- flashcards
- 151
- flashcards
- per day
- ~10 min
- per day
- level
- Beginner · no coding required
- level
- modules
- 8
- modules
What is the Claude Certified Associate — Foundations certification?
Claude Certified Associate — Foundations (CCAO-F) is Anthropic's certification for people who use Claude rather than build on it — operations, marketing, project management, education and communications. The exam guide is unusually direct about this: it targets limited to moderate technical skill and explicitly excludes software developers building APIs or designing agent systems. One of the competencies it tests is recognising the limits of your remit and escalating technical work.
The exam is blueprint-weighted across seven domains, and the balance is lopsided: Output Evaluation and Validation is the single heaviest, followed by Workflow Integration and Solution Design and Governance, Risk and Responsible Use. Read together, those weights say something plain — most of the exam is about judgement, not about clicking the right button. Weights and exam specifics come from the official exam guide and can change; confirm the current figures on Anthropic's pages before you book.
This track covers all seven domains, with card counts mirroring the blueprint so most of your study time lands where the exam does. Every question is a scenario rather than a definition, and every card links to the official Anthropic documentation for the fact it tests — so you can check the source rather than take a flashcard's word for it.
Building on Claude rather than working with it? Claude Certified Developer — Foundations is the technical track, and CCAO-F vs CCDV-F compares the two blueprints — neither exam is a prerequisite for the other.
8 modules, seed to bloom
Each module is a set of flashcards — 151 in total. Answer, review, and watch your knowledge grow from seed to full bloom.
Judging Claude's Output
Recognising when a Claude answer is unreliable — fabricated citations and actions, confident wrong figures, knowledge cutoffs, and the failure modes that survive a skim
16 cardsValidating & Correcting Output
Turning judgement into process — success criteria, sampling, independent ground truth, where the review gate belongs, and leaving evidence that the check actually happened
16 cardsWorkflow Integration & Solution Design
Choosing the right surface and wiring Claude into real work — projects, connectors, skills, artifacts and file handling, plus where the human gate belongs and what a rollout has to standardise
24 cardsGovernance, Risk & Responsible Use
The rules around the work — Usage Policy standards and high-risk obligations, who can see conversations, training and retention defaults, DPAs and BAAs, certifications, enforcement and appeals
23 cardsPrompting & Task Execution
Getting a usable result the first time — stating the brief, supplying the source, chaining long tasks, choosing effort and thinking settings, and recognising when the prompt is not what is wrong
21 cardsProduct & Model Selection
Picking the right Claude for the job — which surface and plan fit the work, what a model tier or a mid-conversation switch actually changes, and how research, plugins, bundles and preview features differ
18 cardsConfiguration & Knowledge Management
Keeping the setup healthy over time — what belongs in project knowledge, how memory and instructions are controlled, exporting data, session security, and how connectors and skills are administered
18 cardsTroubleshooting & Optimisation
Reading the failure correctly — limit and capacity messages, sign-in problems, conversations that outgrow their budget, skills and connectors that stay silent, and knowing when the fault is not yours
15 cardsSample questions
A taste of the real flashcards. Pick an answer, then reveal the explanation.
A researcher asks Claude for sources and gets three URLs, two of which lead to 404 pages. What explains the dead links?
- AClaude produced plausible-looking URLs rather than links it retrieved, a known fabrication failure mode
- BThe pages were fetched correctly but the publishers removed them in the minutes since Claude answered
- CClaude shortened the addresses to save output length, so each link needs expanding before it resolves
- DThe links open only for signed-in subscribers, because sources are served from Claude's own archive
Asked whether it is sure, Claude changes its answer. What has that change actually demonstrated?
- AThat the model reacted to pressure in the prompt, which is no evidence that either answer is right
- BThat the first answer was wrong, since a correct answer would have survived the challenge intact
- CThat the second answer is verified, since re-checking triggers a stricter internal review pass
- DThat confidence was low the first time, which the model reveals by revising when it is questioned
A manager worries that connecting a document store to Claude will expose files their team cannot normally open. What actually governs access?
- AThe user's own permissions in the source system, which the connector cannot exceed on their behalf
- BThe connector's own service account, which reads the whole store and filters results afterwards
- CThe plan tier of the workspace, since higher tiers unlock deeper access into connected systems
- DThe chat's tool access mode, since a broader mode widens what the connector is allowed to read
A firm's risk officer asks whether work done on a commercial Claude plan is used to train Anthropic's models. What is the documented default?
- ACustomer content from the commercial products and the API is not used for model training by default
- BCustomer content is used for training unless an administrator opts the organisation out each year
- CCustomer content is used for training but only after identifying details have been stripped from it
- DCustomer content is used for training only where a conversation has been flagged by a classifier
Learn it once, keep it for good
Answer a question
Each card is one practical concept with multiple options. Pick what you think is right.
Get the full answer
See the correct option plus a clear explanation, and a link to deeper docs when one is available.
Review at the right time
A spaced-repetition engine (SM-2 or FSRS) resurfaces each card just before you would forget it.
Why these foundations are worth your time
Know when not to trust the answer
Fabricated citations, confident wrong figures and claims of actions never taken all read as competent prose. Spotting them is the habit the exam weights most heavily — and the one that protects your work.
Turn judgement into process
Success criteria agreed up front, sampling that means something, and an independent source rather than asking the model whether it is sure. Repeatable review beats a good instinct.
Stay inside the Usage Policy
High-risk uses such as hiring, credit, insurance and legal advice carry human review and disclosure obligations. Knowing which line you are near is a professional skill, not a compliance chore.
Certification-ready
The cards map to all seven CCAO-F domains, weighted to match the blueprint, so the heaviest domains get the most practice.
Common questions
Do I need to be technical to take the CCAO-F exam? +
No. The exam guide targets professionals with limited to moderate technical skill — operations, marketing, project management, education and communications — and explicitly excludes developers building APIs or agent systems. It even tests recognising when a task should be escalated to someone technical.
What is the difference between CCAO-F and CCDV-F? +
CCAO-F is about using Claude well: evaluating output, designing workflows, and governance. CCDV-F is about building on Claude: the Messages API, model selection, agents and tools. Neither is a prerequisite for the other, so you can start with whichever matches your job.
How long does it take? +
About 10 minutes a day. Spaced repetition means short, frequent sessions beat cramming — most learners get comfortable with the core judgement calls within a couple of weeks.
Is it free? +
Yes, completely free. No registration or credit card is required, and all your progress is stored locally in your browser.
Will this prepare me for the CCAO-F exam? +
It covers all seven exam domains, weighted to mirror the official blueprint so most cards land on output evaluation and workflow integration. It is a companion, not a course or an exam dump: pair it with the official Anthropic exam guide and real hands-on use of Claude.
Ready to use Claude with judgement?
Plant your first seed today. Ten minutes a day grows the habits the CCAO-F exam actually rewards.