The Claude Certified Associate – Foundations (exam code CCAO-F) is Anthropic's certification for people who use Claude as a productivity tool at work, with no coding required. The official CCAO-F exam guide describes a 60-question, 120-minute proctored exam that costs $99 and requires no software-development or API experience, and Anthropic's eligibility rules limit it to people who work at organizations in the Claude Partner Network.

It is the least technical of Anthropic's four Claude exams, and Anthropic's July 23, 2026 overview of its role-based certifications marks it as new. Its exam guide (version 1.0, effective July 2026) names operations, marketing, project management, education and communications as the roles it fits, and the Partner Academy certifications page pitches it "for consultants, sellers, and delivery leads who guide customers toward the right Claude use cases." For all four exams side by side, plus the free Claude Academy badges, read our Claude certification guide. This page covers the Associate in depth: what it tests, who can sit it, how to prepare from your own work, and whether it is worth $99.

What is the Claude Certified Associate exam?

The Claude Certified Associate exam tests whether you can apply Claude to business and productivity tasks with minimal guidance: picking the right features and model, structuring prompts, checking output, and escalating technical work to Claude developers and architects. The exam guide describes candidates as having "limited to moderate technical expertise" and places them "between casual AI prompt users and technical AI practitioners."

Anthropic wrote it for people who use Claude as a productivity tool and build Claude Projects in their day-to-day work: staff who run AI-assisted workflows inside a company, and outside consultants who help clients find use cases and redesign processes. The same guide excludes software developers who build against APIs or design agentic systems, lists no mandatory prerequisites or required courses, and says the credential is awarded "based on exam performance alone."

CCAO-F at a glance

multiple-choice and multiple-response questions

120

minutes to answer, about 135 in the seat

720

scaled score to pass, on a 100 to 1,000 range

$99

per attempt, before partner-tier discounts

Source: Anthropic's CCAO-F exam guide and certification FAQ, as of October 2026.

Per the guide, the exam has 60 multiple-choice and multiple-response items, and each item tells you how many answers to pick. You get 120 minutes of testing, and Anthropic's certification FAQ says to plan on about 135 minutes in the seat, including check-in and a short survey. You pass with a scaled score of 720 on a 100 to 1,000 range, the same cut as the other three Claude exams. The exam is closed book, so notes, documentation and AI assistants stay out of reach, and browser translation tools are not permitted. You can sit it online through Pearson VUE's OnVUE proctoring or at a Pearson test center, in English or Japanese.

CCAO-F vs CCAR-F: which exam is which?

CCAO-F and CCAR-F differ by one letter and test different jobs. Pearson VUE's Anthropic page lists CCAO-F as Claude Certified Associate – Foundations and CCAR-F as Claude Certified Architect – Foundations. On the Partner Academy certifications page, Architect – Foundations costs $125 and the Architect track targets "partners who design Claude solutions end to end," while the Associate is the only one of the four exams priced at $99. Anthropic's FAQ also lists the Associate as the one exam that does not count toward a firm's Claude Partner Network eligibility. If you design agent systems for clients, our Claude Certified Architect guide covers CCAR-F.

Who can take the Claude Certified Associate exam?

Only people who work at an organization in the Claude Partner Network can take the Claude Certified Associate exam. Anthropic's eligibility page requires you to sign in to Anthropic Partner Academy with a company email on a domain listed in your organization's partner record, and it warns that "a personal address won't work." Being a paying Claude customer is not enough either: asked whether customers or non-partners can get certified, the certification FAQ answers that certification "is currently available only to organizations in the Claude Partner Network."

You also need to be 18 or older with a government-issued ID. If your company's domain is missing from the partner record, the eligibility page says to email partner-support@anthropic.com, and adding a domain "usually takes 7 to 10 days." Once you pay for the exam on Partner Academy, Anthropic emails you a Pearson VUE booking link.

If your employer is not a partner, the only route runs through the company. Anthropic's Partner Network announcement from March 12, 2026 says "any organization that is bringing Claude to market is eligible to join" and that "membership is free of charge." Firms apply at claude.com/partners. The test is bringing Claude to market. Anthropic's eligibility page tells organizations that deliver Claude for clients to apply, and tells people who administer Claude inside their own company to point colleagues to Claude Academy; their Anthropic account team can record the organization's interest in certification.

Your firm's partner tier sets your price. The certification FAQ says Registered-tier partners pay full price, Select, Preferred and Global Premier partners get 50% off at checkout, and Global Premier partners pay nothing for any certification exam through December 31, 2026. Ask your partnerships team which tier your firm holds before you pay the $99 yourself.

Booking, exam day and retakes

You book through your Pearson VUE account, online or at a test center, and your Pearson profile name must match your government-issued ID exactly. Pearson VUE's Anthropic page says name corrections go to certifications-support@anthropic.com at least 24 hours before the exam, and a mismatch at check-in costs you the fee. The same page says "Exams cannot be canceled or rescheduled less than 48 hours prior to your appointment," while the July exam guide allows changes up to 24 hours before. Plan around the stricter 48 hours. If you need accommodations, the guide says to get Pearson VUE's approval before you schedule.

On exam day, online candidates stay in webcam view for the whole session, and the guide bans phones, smart watches, headphones, study materials and recording devices. You accept a non-disclosure agreement before the first question. Your score report lists the percentage you got right in each domain, and the guide says those domain percentages help you study but do not decide the pass. The certification FAQ says your score shows on screen when you finish. If you fail, it sets waiting periods of 14 days after the first attempt, 30 after the second and 90 after the third, allows at most four attempts per exam in any rolling 12 months, and charges the full fee for each retake (tier discounts still apply).

The Claude Starter System — free PDF kit

Outside the Partner Network? Build the everyday skills anyway

The Claude Starter System is our free PDF: 20 prompts, skills, and five-minute guides for everyday Claude work on Free, Pro, or Max. It is not exam prep and carries no credential. Use it to get finished work out of Claude this week.

Download for free

What's on the CCAO-F exam? The seven domains and weights

The CCAO-F exam covers seven domains, and the heaviest is Output Evaluation and Validation at 21%, while prompting gets 14%. The exam guide's blueprint weights them as follows: Prompting and Task Execution 14%, Output Evaluation and Validation 21%, Product and Model Selection 12%, Workflow Integration and Solution Design 16%, Configuration and Knowledge Management 12%, Governance, Risk, and Responsible Use 15%, and Troubleshooting and Optimization 10%.

Domain	Weight	About how many of 60	What it asks you to do
1. Prompting and Task Execution	14%	8	Write prompts for business tasks, break complex requests into steps, iterate, and adapt to analysis, research, drafting or brainstorming
2. Output Evaluation and Validation	21%	13	Check accuracy and completeness, spot hallucinations and bias, fact-check, decide when a human must review, and pick the right output format
3. Product and Model Selection	12%	7	Choose between chat, Projects, research mode and artifacts; match Haiku, Sonnet or Opus to cost, speed and quality; manage context and memory
4. Workflow Integration and Solution Design	16%	10	Analyze requirements, fit Claude into existing workflows, and explain its value and limits to stakeholders
5. Configuration and Knowledge Management	12%	7	Set up Projects with instructions and knowledge, manage uploads and connectors such as Google Drive and Gmail, and keep them current
6. Governance, Risk, and Responsible Use	15%	9	Separate appropriate from inappropriate uses, apply data sensitivity and privacy rules, and follow your organization's AI policy
7. Troubleshooting and Optimization	10%	6	Diagnose weak prompts and outputs, adjust from feedback, and make workflows more efficient

The question counts in the third column are our arithmetic, not Anthropic's: each weight times 60, rounded. The guide says the percentages show the approximate share of scored items per domain, so use them as a study-time budget. Add up Output Evaluation (21%), Governance (15%) and Troubleshooting (10%) and you get 46% of the exam, all about judgment calls: checking what Claude produces, deciding what data and tasks belong in Claude, and fixing weak results. If you write good prompts but skip those checks, nearly half the blueprint aims at your weak spot.

The guide's three sample questions show the style: a short workplace scenario, four options, one best action. In the first, Claude writes a confident summary of a new regulation that cites a specific subsection, and the right move is to check that subsection against the official text before the summary reaches the compliance team; asking Claude to rate its own confidence is a wrong answer. The second asks which model fits a high volume of short customer replies where speed and cost matter more than deep reasoning, and the answer is a faster, lower-cost model. In the third, a project manager wants to upload a spreadsheet of customer names and account numbers under a policy that restricts regulated personal data, and the answer is to remove or anonymize the identifiers first. None of the three needs technical knowledge; each tests a judgment call: verify before you share, match the model to the job, and protect regulated data.

The July guide's model domain names Haiku, Sonnet and Opus. Anthropic's current model lineup also includes Fable, which it describes as built "for demanding reasoning and long-horizon agentic work." Anthropic's model selection guide frames the choice as "balancing capabilities, speed, and cost" and lists "high-volume, straightforward tasks" among the cases for starting with Haiku, the same logic as sample question two. Pearson VUE says the Associate exam gets updated content starting December 7, 2026, with updated exam guides on November 17, 2026, and tells anyone testing on December 7 or later to prepare from the updated guide. For model choice in everyday work, see our comparison of Sonnet vs Opus vs Haiku.

How do you prepare for the Claude Certified Associate without coding?

Prepare for the Claude Certified Associate by doing your real work in Claude with the blueprint open, then filling gaps with Anthropic's free courses. The exam guide says "there is no single required course" and that Anthropic does not guarantee any resource will get you a pass. Its advice: study the blueprint and self-assess, read Anthropic's documentation and help articles on features such as Projects, Artifacts, Memory, Skills and code execution, practice structuring and iterating prompts, build a real Project with instructions and knowledge sources, practice responsible-use judgment, and work through the sample questions.

If your firm is a partner, Anthropic Partner Academy has a free Claude Certified Associate – Foundations Prep Course with eight modules, one per domain plus an 8-minute summary, about six and a half hours in total. Its longest module, Evaluating & Validating Claude's Output at 74 minutes, lines up with the heaviest domain. Pearson VUE's Anthropic page says this prep training "is available to members of the Claude Partner Network."

The detail that helps everyone else sits on that same prep course page. Before you start it, Anthropic recommends three courses: Claude 101, AI Fluency: Framework & Foundations, and AI Capabilities and Limitations. All three are on Claude Academy, Anthropic's free public course site at academy.claude.com (Anthropic runs it; it has no connection to AItomation Academy). The Claude Academy catalog lists Claude 101 at 13 lessons and 2.5 hours, AI Fluency: Framework and foundations at 14 lessons and 4 hours, and AI capabilities and limitations at 13 lessons and 3.5 hours, each with a completion badge. That is about 10 hours of Anthropic-made material you can take without a partner email. Add Introduction to Claude Cowork (14 lessons, 2.5 hours), which also sits on Partner Academy's Claude Associate learning path.

Courses teach the vocabulary, and the exam's scenarios are workplace situations. This drill plan maps each domain to something you can do in your own job this month, with a test for when you are ready:

Domain	Real-work drill	You are ready when
Prompting (14%)	Rewrite one recurring request from your week as a structured prompt with role, context, task and output format. Run it, then improve it once.	You can split a messy request into steps without notes.
Output evaluation (21%)	For one week, list every number, name, date and citation in Claude's drafts and check each against its source before you send anything.	You catch an invented detail before a reviewer does.
Product and model selection (12%)	Run one task in plain chat, in a Project and with Research. Run a high-volume task on Haiku and a hard one on a larger model.	You can say which feature and model fit a task, and why.
Workflow integration (16%)	Map one recurring process step by step and mark which steps Claude takes and which stay with a person.	You can explain Claude's value and limits to a skeptical stakeholder in five sentences.
Configuration and knowledge (12%)	Build one Project with instructions and three to five knowledge files, connect one tool, and update both after a week.	The Project gives consistent output without re-pasting context.
Governance (15%)	Read your company's AI policy and sort ten real tasks into fine, anonymize first, or keep out of Claude.	You can name the policy reason behind each call.
Troubleshooting (10%)	Log five bad outputs with the cause (missing context, vague instruction, wrong model or feature) and the fix.	You fix a weak output by changing the input instead of regenerating.

You can run almost all of these drills on Claude's Free plan. Anthropic's pricing page lists Projects (up to five), Artifacts, memory, Skills, code execution, connectors and web search on Free, and the Projects help article confirms free accounts get Projects with their own instructions and knowledge base. Two items need a paid plan. Research runs only on Pro, Max, Team or Enterprise, and the Free plan includes only the Sonnet and Haiku models, so Opus starts at Pro ($20 a month, or $17 a month billed annually). If you are weighing paid tiers, our Claude Pro vs Max comparison covers the trade-off.

For the prompting domain, our Claude prompting guide walks through structured prompts with worked examples. For Anthropic's free and paid training side by side, see our breakdown of Anthropic courses, and for options beyond Anthropic, our list of the best Claude courses. To quiz yourself against the blueprint with scenarios from your own job, paste one domain from the guide into this prompt:

I am preparing for the Claude Certified Associate – Foundations
exam (CCAO-F). Quiz me on this domain:
[paste one domain name and its task statements from the
official exam guide]

My job: [role, industry, the tasks I use Claude for]

Rules:
1. Write 5 scenario questions set in my kind of work. Each has
   4 options and says whether to pick 1 answer or several.
2. Ask one question at a time and wait for my answer.
3. After each answer, tell me if I was right, why the best
   option is best, and why each other option is weaker.
4. Favor judgment calls: verifying output, protecting sensitive
   data, choosing a feature or model, knowing when to escalate.
5. Do not present these as real exam questions. If you are
   unsure how a Claude feature works, say so instead of guessing.

Treat Claude's questions as practice, not the real item bank. When an explanation names a feature or a plan detail, check it against Anthropic's help center before you believe it, which doubles as practice for the 21% output-evaluation domain. For that domain in your real work, run this review before any Claude draft leaves your hands:

Before I send this draft, act as my reviewer.

1. List every factual claim: numbers, names, dates, quotes,
   citations, product or policy details.
2. Mark each one: (a) supported by the sources I gave you,
   (b) not in my sources and needs checking, or (c) likely wrong.
3. Flag any wording that sounds more certain than the sources.
4. Flag any personal or confidential data that should not leave
   my team.
5. Say who should review this before it goes out, if anyone.

Draft: [paste]
Sources: [paste or attach]

Marko Sudar — founder of AItomation Academy

Want feedback on your Claude workflows every week?

1,300+ non-technical professionals learning together
Claude, Claude Cowork & Claude Code training
The full 550+ prompt & skill Vault included with Premium
A weekly live call to get unstuck

Join the Academy

Is the Claude Certified Associate worth it?

The Claude Certified Associate is worth $99 if you work at a Claude partner firm and your job involves guiding clients or colleagues on Claude, because the exam is short, cheap, and tests work you already do. If your organization is outside the Claude Partner Network, it is not worth chasing, because Anthropic's eligibility rules won't let you register. Run these three checks before you book:

Should you book the Associate? Three checks

Is your organization in the Claude Partner Network?

No: you cannot register. Use the blueprint as a free study plan and build proof from real work. If you run a firm that brings Claude to market, network membership is free.

Do you advise clients or teammates on how to use Claude?

Yes: you are the audience Anthropic names, from consultants and delivery leads to project managers. No: the blueprint still works as a self-assessment, and the skills matter more than the badge.

Will you renew it every 12 months?

Per the exam guide, on-time renewal is a free, non-proctored assessment. Let it lapse and you pay for the full exam again.

Two facts shape the value for your employer. The Associate does not count toward a firm's Partner Network tier, per Anthropic's certification FAQ, so your firm gains skills from it but no tier credit. And the credential expires: the exam guide says it is valid for 12 months, on-time renewal is a free, non-proctored assessment on Partner Academy, a lapse means retaking the full exam at the full fee, and Anthropic may require a full retake if the exam content changes significantly. When you pass, the FAQ says you get a Credly badge you can share on LinkedIn and a PDF certificate you can download.

If you are outside the network, put the hours into the skills. Our how to learn Claude roadmap lays out a four-week learning path, and the portfolio checklist in our Claude certification guide shows how to turn real tasks into proof a hiring manager can check. If you sell Claude work to clients, our guide to AI for consultants covers the workflows that clients pay for.

Frequently Asked Questions

Is the Claude Certified Associate the same as a Claude Academy badge?

No. Claude Academy courses are free, and Anthropic's eligibility page says they "end with a quiz, and passing earns a certificate of completion you can share." The same page calls a certificate of completion and a certification "two different things." The Associate is a proctored Pearson VUE exam with a scaled passing score, and Anthropic's certification FAQ says passing earns the credential and a Credly badge.

Do I need to know Claude Code or the API for the Associate exam?

No. The exam guide says no software-development or API experience is needed, and the July 2026 version never mentions Claude Code. Its blueprint names chat, Projects, research mode, artifacts, connectors and the Haiku, Sonnet and Opus models. If you want Claude Code for other reasons, our Claude Code pricing guide explains which plans include it.

Are there official practice exams for CCAO-F?

No full practice exam. Anthropic's certification FAQ says the practice exam on its previous platform was retired in the move to Pearson, and it points candidates to the sample questions in each exam guide instead. The Associate guide has three, so any CCAO-F practice test you find for sale comes from a third party, not Anthropic.

How long does it take to prepare for the Claude Certified Associate?

The exam guide gives no study-time estimate. The official material adds up to roughly 16.5 hours: about 10 hours for the three recommended courses on Claude Academy and about 6.5 hours for the partner prep course. If you already use Claude every day at work, the drills above count for more than the course hours.

What happens to my certification after the December 7 exam update?

Your credential stays valid for 12 months from the date it was awarded, but the exam guide warns that if exam content changes significantly, Anthropic may require holders to retake the full exam instead of the free renewal assessment. Pearson VUE says the updated content starts December 7, 2026, with new exam guides on November 17.

Claude Certified Associate: CCAO-F Exam Guide

Get The Claude Starter System