The Claude Certified Associate (Foundations) exam, code CCAO-F, is the programme's non-developer credential and its most accessible entry point. No API, no code, no software-development experience. It tests whether you can put Claude to work on real business and productivity tasks: prompting well, judging whether the output is good enough, choosing the right product feature and model, configuring Claude Projects, and using AI responsibly inside an organisation.
This site covers it with 30 lessons, one for each objective in the exam guide, plus a free diagnostic and six full-length mock exams. Below is the published blueprint, so you know exactly what the exam covers.
Who it's for
The Associate exam sits deliberately between two groups: casual users who just type prompts, and technical practitioners who build with the API. It's for professionals who use Claude as a serious productivity tool and build Claude Projects, in roles like operations, marketing, project management, education, communications, and general knowledge work. Internal staff and external consultants both fit.
The guide describes the expected technical level as limited to moderate. What it explicitly excludes is telling: this is not the exam for software developers building against APIs or designing agentic systems, nor for ML, software-engineering, or advanced-AI specialists. That scope belongs to the Developer and Architect credentials. In fact, one of the skills the Associate exam tests is recognising the limits of your remit and escalating complex or technical work to Claude Architects and Developers.
There are no mandatory prerequisites and no required courses. Recommended background is simply regular, hands-on Claude use in a professional setting. The credential is awarded on exam performance alone.
The seven domains
The weighting makes the exam's priorities obvious: it cares most about whether you can tell good output from bad.
| Domain | Weight | Focus |
|---|---|---|
| Prompting and Task Execution | 14% | Writing effective prompts and getting tasks done |
| Output Evaluation and Validation | 21% | Judging quality, accuracy, and whether output is fit to use |
| Product and Model Selection | 12% | Choosing the right feature and the right model tier |
| Workflow Integration and Solution Design | 16% | Fitting Claude into real processes |
| Configuration and Knowledge Management | 12% | Setting up Projects and managing knowledge |
| Governance, Risk, and Responsible Use | 15% | Using AI safely and responsibly at work |
| Troubleshooting and Optimisation | 10% | Diagnosing and improving results |
Output Evaluation and Validation is the heaviest domain at 21%, more than a fifth of the exam. That's the crux of the Associate role: the value isn't in generating text, it's in knowing whether the text is right, complete, and safe to send. Governance, Risk, and Responsible Use is weighted heavily too, at 15%, reflecting how much of workplace AI use is about doing it appropriately, not just effectively.
Product and Model Selection is worth calling out because it's concrete. It covers the model tiers (Haiku, Sonnet, and Opus) and the product features you'd actually reach for, such as Projects, research mode, chat, and artifacts. Knowing which tool fits which job is directly testable.
Where it sits among the four certifications
Associate is the business-user credential in a programme of four Claude certifications. The clean way to think about the three role-based tracks:
- Associate uses Claude as a productivity tool and escalates technical work upward.
- Developer builds and ships Claude-powered applications through the API, Agent SDK, and MCP.
- Architect designs those systems end to end.
At 60 items, 120 minutes, a scaled 720 out of 1,000 to pass, and a 12-month validity, the mechanics match the rest of the programme.
How to use the prep track
Take the free diagnostic first. It shows you where you're weakest before you spend an evening on the wrong domain. Then go through the curriculum in blueprint order, and don't rush Output Evaluation and Validation, which is 21% of the exam on its own. Every lesson is free. Once the practice questions stop catching you out, sit a mock exam: 60 questions in 120 minutes, scored against the real 720 pass mark.
Because this exam is non-technical, keep practising with Claude itself alongside the lessons. Draft real work, then judge it hard: is each output accurate, complete, and appropriate to send? Practise choosing between model tiers for different tasks, and get comfortable setting up a Claude Project with the right knowledge and instructions.
This is an independent community resource, not affiliated with or endorsed by Anthropic. Registration for the official exam runs through Anthropic's Partner Academy, delivered by Pearson VUE.