Community Study Guide · v1.1 · September 2026
Claude Certified
Associate: Foundations
The non-technical Claude credential — for people who use Claude
to get real work done. No coding required. Free to share.
| 7Exam Domains |
60Questions |
$99Exam Fee |
~20hStudy Content |
The core idea:
You don't need to write code to earn this. You need to use Claude well, judge its output critically, and know when to validate, escalate, or hand off to a developer.
promptgoblins.ai · A community for people who build with AI
About This Guide
This guide was built by Prompt Goblins. The Claude Certified Associate – Foundations (CCAO-F) is the entry point of Anthropic's four-credential program and the only one aimed at non-technical professionals — people in operations, marketing, project management, education, HR, communications, and consulting who use Claude as a day-to-day productivity tool.
If you build applications, agents, or integrations against the API, this is not your exam — you want the Developer or Architect credentials, covered in our companion Builder Track guide. The Associate exam explicitly stops where code begins: one of its own skills is recognizing when to escalate technical work to Architects and Developers.
Access: Like all four Claude exams, the Associate is currently available only to organizations in the Claude Partner Network (free to join for any organization bringing Claude to market), delivered through Pearson VUE. No software-development or API experience is required to sit it.
Jump to: Domains · Study Plan · Cliffs Notes · Exam Logistics · Change Log
Who should take it
- Business analysts, project/operations managers, marketing & communications professionals, HR and education staff, and consultants
- People who already use Claude regularly at work — building Projects, drafting, analyzing, researching
- Anyone who wants a credential proving they can apply Claude responsibly and effectively without pretending to be an engineer
Recommended (not required): regular hands-on Claude use in a professional setting; comfort with structured problem-solving and workflow design; a practical understanding of AI limits (hallucinations, context limits, data sensitivity). There are no mandatory prerequisites — the credential is awarded on exam performance alone.
How to read the attribution tags
| OFFICIAL | On an Anthropic-curated course list. |
| ANTHROPIC | Anthropic-published, mapped to exam domains. Academy courses and official help/docs. |
| SUPPLEMENTAL | Community / third-party. Items tagged optional can be skipped without hurting coverage. |
Tags are deliberately grayscale — color always means an exam domain, never a source tier.
Exam Domains at a Glance
Seven domains. Note where the weight sits: Output Evaluation & Validation (21%) is the single largest — this exam cares most about whether you can judge Claude's output, not just prompt it.
21%Output Evaluation & Validation
Judge accuracy & completeness, spot hallucinations/bias, fact-check, decide when human review is needed, adapt outputs for the audience.
16%Workflow Integration & Solution Design
Use Claude for research, planning, and process optimization; integrate it into existing workflows; communicate its value and limits.
15%Governance, Risk & Responsible Use
Appropriate vs. inappropriate use cases; data sensitivity, privacy, regulation; organizational AI policy; ethics.
14%Prompting & Task Execution
Effective prompts, task decomposition, iteration, adapting strategy by task type (analysis, research, drafting, brainstorming).
12%Product & Model Selection
Pick the right feature (Projects, research mode, chat, Artifacts); Haiku vs. Sonnet vs. Opus; manage context & memory limits.
12%Configuration & Knowledge Management
Configure Projects with instructions & knowledge sources; manage connectors (Drive, Gmail); write system-level instructions; keep them current.
10%Troubleshooting & Optimization
Diagnose weak prompts/outputs, adjust based on results, optimize workflows for efficiency.
Study Plan
This is a short track — most candidates are ready in ~17 hours of study (~20h if you add the two optional rows, 5 and 9) plus their existing hands-on experience. Work top to bottom.
Fast path: short on time? Prioritize AI Fluency: Framework & Foundations, AI capabilities and limitations, and the hands-on Project units — those map most directly onto Prompting and Output Evaluation & Validation, the two heaviest domains. Introduction to Claude Tag is optional and only worth your time if your team uses Claude in Slack.
Migrating from Skilljar: Anthropic moved its free courses from anthropic.skilljar.com to Claude Academy (academy.claude.com) around August 20, 2026. The old Skilljar pages still load but are deprecated — use the Claude Academy links below. Progress does not auto-transfer: academy.claude.com shows an import banner only if your Claude account email matches the email you used on Skilljar, and that's rolling out gradually. Certificates you already downloaded as PDFs from Skilljar remain valid.
| # | Resource | Platform | Format | Time | Domains | ✓ |
|---|
| 1 | Claude 101 O | Claude Academy | Interactive | 2.5h | Product, Config | ☐ |
| 2 | AI Fluency: Framework & Foundations O | Claude Academy | Course | 4h | Prompting, Validation, Governance | ☐ |
| 3 | AI capabilities and limitations A | Claude Academy | Interactive | 3.5h | Validation / output evaluation (21%): next-token prediction, knowledge limits, working memory, steerability | ☐ |
| 4 | Introduction to Claude Cowork A | Claude Academy | Interactive | 2.5h | Workflow, Product | ☐ |
| 5 | Introduction to Claude Tag A optional | Claude Academy | Interactive | 2.5h | Workflow — only if your team uses Claude in Slack | ☐ |
| 6 | Claude features: Projects, Artifacts, Memory, connectors A | Claude Academy | Reading | 1.5h | Config, Product | ☐ |
| 7 | Responsible use & the 4D "Discernment/Diligence" ideas A | Claude Academy | Reading | 1h | Governance, Validation | ☐ |
| 8 | Hands-on: configure a Project & evaluate its output S | Claude.ai | Practice | 1.5h | All | ☐ |
| 9 | Community practice questions S optional | CertSafari / others | MCQs | 1h | All | ☐ |
Most Associate community practice material is written for the Architect exam and is too technical for this credential — use it only to practice question style, not content. There is no official Anthropic practice exam (it was retired in the move to Pearson VUE).
Team leads rolling Claude out beyond individual use may also want Building effective human-agent teams (beta) (Claude Academy, 45 min) — not required for this exam.
Concept Cliffs Notes
Evaluating output is the whole game (21%)
Self-reported confidence is not evidence. If Claude cites a specific figure, subsection, or statistic — especially for a compliance or finance audience — verify it against the source before you rely on it. Confident-sounding output can be a hallucination.
Know when a human must review: regulated content, external-facing claims, anything with legal, financial, or safety consequences. "Edit, adapt, refine, and compare" outputs for the intended audience rather than shipping the first draft.
Responsible & safe use (15%)
Data sensitivity first. Before uploading a spreadsheet with names and account numbers, remove or anonymize the personal identifiers to match policy — don't upload as-is and don't rely on telling Claude "not to retain" it. Follow your organization's AI policy and privacy rules.
Right use case: know what Claude is and isn't appropriate for, and escalate technical or high-risk work rather than forcing it.
Picking the right product & model (12%)
Match the model to the task: a faster, lower-cost tier (Haiku) for straightforward high-volume work like short reply drafts; a top tier (Opus) for complex reasoning. Using the most capable model for everything just wastes cost and time.
Pick the right feature: Projects for persistent instructions + knowledge; Artifacts for produced documents; research mode for multi-source work; plain chat for quick tasks. Manage context limits — know when to restart, summarize, or persist.
Prompting & iteration (14%)
Decompose then iterate. Break a complex request into structured steps, and refine the prompt based on what comes back. Adapt your approach to the task type — analysis, research, drafting, and brainstorming each want a different framing.
Exam Logistics
| Exam code | CCAO-F |
| Questions | 60 items (multiple-choice & multiple-response; each item says how many to select) |
| Duration | 120 minutes |
| Delivery | Pearson VUE — online proctored or test center |
| Proctoring | Closed-book; no notes, AI assistants, or external docs |
| Scoring | 100–1,000 scale · Pass at 720; percent-correct per domain reported |
| Cost | $99/attempt (partner-tier discounts may apply) |
| Retake | Up to 4 attempts per rolling 12 months; 14/30/90-day waits after attempts 1/2/3 |
| Validity | 12 months; renewable via a free non-proctored assessment |
| Badge | Digital credential via Credly, shareable on LinkedIn |
| Eligibility | Claude Partner Network organizations (network is free to join) |
| Where it sits | Entry tier. Escalate technical work to Developer ($125) and Architect ($125 / $175) credentials. |
Sample item (illustrative, from the official guide). A project manager wants to upload a spreadsheet of customer names and account numbers so Claude can analyze trends, but policy restricts sharing regulated personal data. The right move: remove or anonymize the identifiers before uploading, consistent with policy — not upload as-is, not "tell Claude not to retain it," and not skip the analysis entirely.
Change Log
v1.1 — September 26, 2026
Claude Academy migration — Anthropic's free courses moved from Skilljar to academy.claude.com; exam and registration unchanged.
- Platform migration: Anthropic's free courses moved from anthropic.skilljar.com to Claude Academy (academy.claude.com) around August 20, 2026. Same courses, but the new site states longer lesson counts and durations (corrected below); certification registration and delivery (Pearson VUE, via the Claude Partner Network) has not moved and stays on anthropic-partners.skilljar.com.
- Duration corrections: Claude 101 1h → 2.5h; AI Fluency: Framework & Foundations 1h → 4h; Introduction to Claude Cowork 1h → 2.5h.
- New study-plan units: added AI capabilities and limitations (3.5h, Claude Academy) directly targeting the Output Evaluation & Validation domain, and Introduction to Claude Tag (2.5h, optional, Slack teams).
- Study-plan total: ~8h → ~17h core (~20h with optionals), reflecting the corrected durations and the two added units.
- Link fix: the "Claude features" reading now points at the Claude product page on Claude Academy (anthropic.com/learn redirects there).
- Link fix: CertSafari link corrected to its Foundations-specific page.
v1.0 — July 2026
Initial release. Blueprint, domain weights, exam format, logistics, and sample questions taken directly from Anthropic's official Claude Certified Associate – Foundations Exam Guide v1.0 ("Effective July 2026," code CCAO-F), cross-checked against the Anthropic certification FAQ and Pearson VUE, on July 19–20, 2026. Course links verified live; all documentation links use current canonical domains. Supplemental items are marked optional and kept in study order.
Claude Certified Associate – Foundations Study Guide · v1.1 · September 2026 · promptgoblins.ai · A community for people who build with AI
Blueprint summarized from Anthropic's official CCAO-F exam guide v1.0. Free to share. Discuss & get updates: promptgoblins.ai/t/86
