CCAO-F · Claude Certified Associate – Foundations: Preparation Track¶
Unofficial study material. This track is written independently. It is not endorsed by Anthropic and contains no live exam content. Exam facts below come from
registry.yaml, which was checked against the official exam guide v1.0 (effective July 2026) on 2026-10-24. The human-readable summary is exam-facts.md. Always read the official guide on the Anthropic Partner Academy certifications page before you book.
What the credential is for¶
The Claude Certified Associate – Foundations (exam code CCAO-F) certifies that you can use Claude to get real business work done with little hand-holding. In practice that means five things:
- Getting structured business tasks done with Claude without step-by-step guidance.
- Using the built-in features of the Claude apps (Projects, Artifacts, research, connectors, memory, Skills) to streamline a workflow.
- Spotting where a process could be improved with Claude.
- Picking an approach that balances quality, speed and cost.
- Knowing Claude's limits, and handing complex or technical work to people with Claude Architect or Developer skills.
It is the entry-level credential in the Claude Certification Program. It is about judgement while using the Claude apps. It does not test coding, the API or agent architecture.
Who it is for¶
In the official guide's terms, the target candidate relies on Claude day to day to get work done and builds Claude Projects, typically in an operations, marketing, project-management, teaching, communications or other knowledge-work role. That includes internal staff who run AI-assisted workflows day to day and external consultants who help clients find use cases and redesign processes. The Partner Academy page also pitches it to consultants, sellers and delivery leads in partner organisations.
It is not aimed at software developers building against the API, ML specialists or people designing enterprise AI architecture. Those topics belong to the Architect and Developer credentials. If you are an engineer, this track is still a good warm-up for CCAR-P, because the judgement it teaches (verify outputs, protect data, choose the right tool) carries over.
The minimally qualified candidate¶
Exam items are pitched at a minimally qualified candidate: someone who just meets the bar. Picture a person who:
- uses Claude regularly at work, not just for one-off questions;
- structures prompts deliberately and breaks big jobs into steps;
- is comfortable with Projects, Artifacts and similar features, and has moved past simple question-and-answer use towards redesigning processes;
- understands where AI creates value in their organisation, where adoption risks sit, and how responsible-use practices apply;
- knows the practical limits of AI: hallucinations, context constraints and data sensitivity.
If that description fits you on most days, you are close. If several parts sound unfamiliar, plan for the longer study sequence below.
Format and logistics¶
All figures come from registry.yaml (verified 2026-10-24).
| Item | Detail |
|---|---|
| Items | 60, multiple-choice and multiple-response (each item says how many answers to pick) |
| Time | 120 minutes (Pearson suggests planning for about 135 minutes of seat time, including check-in) |
| Delivery | Pearson VUE: online proctored (OnVUE) or at a Pearson test centre |
| Passing standard | Scaled score of 720 on a 100–1,000 scale, criterion-referenced (you are measured against a fixed standard, not other candidates) |
| Exam fee | $99 USD list price; the checkout price can reflect a partner-tier discount |
| Results | Pass/fail plus scaled score, and percent-correct by domain on the score report |
| Validity | 12 months from the date the credential is awarded |
| Prerequisites | None mandatory. Regular hands-on Claude use is recommended; no coding or API experience is needed |
The per-domain percentages on your score report help you see weak areas, but your pass or fail result depends only on the total scaled score.
Registration and eligibility¶
- Who can sit it. Certification is open to people at organisations in the Claude Partner Network. You register with a partner email address on a recognised company domain; personal email addresses are not accepted. Candidates must be at least 18. Customers and non-partners cannot certify at the time of writing (as of 2026-10-24). The free Claude Academy courses are open to everyone but award completion certificates, not certifications.
- Partner tier. The Associate credential does not count toward Claude Partner Network tier eligibility; the Developer and Architect credentials do.
- How to register. Go to the CCAO-F certification page on the Anthropic Partner Academy, download the exam guide, read the Certification Terms and Conditions and the Exam Policy, register and pay, then create your Pearson VUE account and schedule either online proctoring or a test centre.
- Name match. The name on your Pearson profile must match your government-issued photo ID exactly. Corrections go to Anthropic certification support before you schedule (see exam-day.md).
Retakes and renewal¶
- Retakes. Wait 14 days after a first failed attempt, 30 days after a second and 90 days after a third. You may attempt the same exam at most four times in a rolling 12 months. Each attempt costs the full fee.
- Renewal. The credential lasts 12 months. Renewing before expiry costs nothing: you catch up on what has changed, then pass an unproctored check on the Anthropic Partner Academy. Let it expire and you sit (and pay for) the whole exam again. Anthropic may require a full retake if exam content changes significantly.
Upcoming change: updated guide on 2026-11-17¶
Pearson VUE states that updated exam guides will be published on 2026-11-17 and that CCAO-F will carry updated content from 2026-12-07. If you sit the exam on or after 2026-12-07, prepare from the updated guide. This track will be re-checked after the update; see the verification log. The skills taught here (prompting, verification, product choice, governance) are unlikely to change, but feature names, plan availability and model names may.
How to use this track¶
| Step | Use | Why |
|---|---|---|
| 1 | Read this page and readiness-assessment.md | Rate yourself 1–5 on every objective before you start |
| 2 | Work through study-guide/domain-1.md … domain-7.md |
The detailed notes, decision rules and mini-scenarios |
| 3 | Do the matching exercises in practical-exercises.md | CCAO-F rewards hands-on habit, not memorised definitions |
| 4 | Read decision-scenarios.md and common-misconceptions.md | Practise the "best next action" judgement the exam measures |
| 5 | Drill with practice questions and the CCAO-F flashcards | Spaced repetition of facts and rules |
| 6 | Take mock exams under time pressure | Pacing for 60 items in 120 minutes |
| 7 | Final week: revision-notes.md, then exam-day.md | Consolidate, then handle logistics calmly |
The module pages linked from each study guide are written for a mixed audience. Where a module includes Python or API material, you can skip the code: CCAO-F only needs the concepts. The labs are Python-based; if you do not code, use the no-code equivalent in practical-exercises.md instead.
Domains¶
| # | Domain | Weight | Study guide | Key modules |
|---|---|---|---|---|
| 1 | Prompting and Task Execution | 14% | domain-1.md | Clear instructions, Task decomposition, Iterating |
| 2 | Output Evaluation and Validation | 21% | domain-2.md | Hallucinations, Grounding and citations, Human oversight |
| 3 | Product and Model Selection | 12% | domain-3.md | Product ecosystem, Model families, Context windows |
| 4 | Workflow Integration and Solution Design | 16% | domain-4.md | Solution design process, Workflow patterns, Stakeholder communication |
| 5 | Configuration and Knowledge Management | 12% | domain-5.md | Projects and knowledge, Connectors, System prompts |
| 6 | Governance, Risk, and Responsible Use | 15% | domain-6.md | Responsible use, Data disclosure, AI governance |
| 7 | Troubleshooting and Optimization | 10% | domain-7.md | Diagnosing failures, Prompts and outputs troubleshooting, Cost optimisation |
Domain 2 carries the most weight (21%). Domains 2, 4 and 6 together make up just over half the exam, and all three are about judgement: is this output right, does this fit the workflow, is this use acceptable?
Recommended study sequence¶
The order below follows dependencies, not domain numbers.
| Phase | Focus | Suggested time |
|---|---|---|
| 1. Foundations | What LLMs are, hallucinations, context windows, how Claude works | 2–3 days |
| 2. Prompting (D1) | Domain 1 study guide plus exercises 1.1–1.3 | 2–3 days |
| 3. Products and models (D3, D5) | Domains 3 and 5 together: build a real Project with instructions and knowledge | 3–4 days |
| 4. Evaluation (D2) | Domain 2, the heaviest domain. Practise fact-checking on your own outputs every day | 3–4 days |
| 5. Responsible use (D6) | Domain 6 plus redaction exercise | 2 days |
| 6. Workflows (D4, D7) | Domains 4 and 7: redesign one real workflow end to end and troubleshoot it | 3 days |
| 7. Consolidation | Mock exams, misconceptions, revision notes, readiness check | 3–5 days |
That comes to roughly three to four weeks at one to two hours a day. If you already use Claude daily at work, the 15-day plan or even the 7-day accelerated plan may be enough. Newcomers to AI should follow the 30-day plan.
Coverage¶
The coverage matrix is generated from objective-map.yaml and the question banks. It shows, for every objective, which module pages, labs and practice questions cover it. A one-page summary is in the CCAO-F cheat sheet.
Product facts used in this track¶
Claude app features change often. Product statements in this track were checked against support.claude.com on 2026-10-24. Plan availability (Free, Pro, Max, Team, Enterprise), feature names and model names can change at any time, so check the Help Center before relying on a detail at work.