60 questions in 120 minutes, scored out of 1000 with 720 to pass. Short business scenarios, with multiple choice, select-TWO and ordering questions (put the steps of a process in the right order). Output Evaluation and Validation is the largest domain at 21%. Time limit, question count and pass mark are as widely reported for exam guide v1.0; confirm on Anthropic's certification pages before you book.
The domains and their weights
Prompting and Task Execution 14%
- Give Claude the context, audience, format and constraints up front; vague prompts get generic answers.
- Show an example of what good looks like when tone or format matters.
- Break big jobs into steps and check each step before moving on.
- Match the approach to the task: brainstorming needs room, summaries need tight limits.
Practise 13 Prompting and Task Execution questions →
Output Evaluation and Validation 21%
- Check facts, figures, names and citations against the source before anything leaves your hands.
- Watch for confident detail the source does not contain.
- Ask Claude to quote the passage behind each claim so you can verify it quickly.
- The person who sends the work is accountable for it, not the tool.
Practise 19 Output Evaluation and Validation questions →
Product and Model Selection 12%
- Larger models suit complex reasoning and careful writing; smaller, faster ones suit high-volume simple tasks.
- Projects keep reusable instructions and files together for repeated work.
- Artifacts suit documents, tables and code you want to edit and reuse.
- Connectors bring company tools and data into the conversation where your organisation allows it.
Practise 11 Product and Model Selection questions →
Workflow Integration and Solution Design 16%
- Start with a repetitive, well-understood task and keep a human review point.
- Decide which steps Claude drafts and which a person approves.
- Write the process down so others can repeat it.
- Measure the result against how the work was done before.
Practise 14 Workflow Integration and Solution Design questions →
Configuration and Knowledge Management 12%
- Use Projects and custom instructions so the team does not repeat context in every chat.
- Keep shared reference files current and remove outdated ones.
- Give each body of knowledge a clear owner.
- Start a fresh conversation when an old one is cluttered.
Practise 11 Configuration and Knowledge Management questions →
Governance, Risk, and Responsible Use 15%
- Follow your organisation's policy on what data may be shared with AI tools.
- Remove or anonymise personal and confidential details you are not permitted to share.
- Keep a person accountable for decisions that affect people, money or reputation.
- Check outputs for bias, and be open about AI use where it matters.
Practise 13 Governance, Risk, and Responsible Use questions →
Troubleshooting and Optimization 10%
- Inconsistent output usually means missing context, examples or constraints.
- Split an overloaded prompt into smaller steps.
- Ask Claude what information it needs before it starts.
- Compare versions side by side and keep the prompt that works.
Practise 9 Troubleshooting and Optimization questions →
Further practice
For a second set of questions written by someone who has sat the exam, Matthew Purcell publishes an independent Associate – Foundations (CCAO-F) practice set with answers and rationales.
Try three sample questions
Pick your answer, then open the reveal. The full bank has 90 questions.
- A checklist for our 40-room hotel: turnover cleans by 2pm, eco linen policy, room-by-room order
- A better, more detailed and more professional checklist for the whole of our housekeeping team to use
- A longer checklist written as if by an expert hotel manager with twenty years of experience
- Three different checklists in different styles so that the team can choose the best one
Reveal answer
- Keep it, as Claude would only include a figure if it were correct
- Round it to 20% so it sounds less precise and is safer to share
- Find the source for the figure or remove it before the update goes out
- Add the words "approximately" before the figure and send it as written
Reveal answer
- Paste the full file in so Claude has all the background it needs
- Check company policy on using personal data with AI tools first
- Remove the employee's name but paste the rest of the file in
- Ask Claude whether it is allowed to see the employee's file
Reveal answer
A four-week study plan
- Week 1, prompting. Lessons in Level 1, then practise Prompting and Task Execution.
- Week 2, checking outputs. The largest domain: practise Output Evaluation until you score 80% or more.
- Week 3, products, workflows and knowledge. Practise the product, workflow and configuration domains.
- Week 4, governance and mock exams. Practise governance and troubleshooting, then take the timed mock until you score 800+.
Frequently asked
Who is the Claude Certified Associate exam for?
Professionals who use Claude in their work, such as operations, marketing, HR, finance and legal teams. It tests good judgement with AI rather than coding.
How long is the exam and what is the pass mark?
It is widely reported as 60 questions in 120 minutes, scored from 100 to 1000 with 720 to pass. Check Anthropic's certification pages for the current details.
What kinds of questions are there?
Short business scenarios with multiple choice, multiple response (select TWO) and ordering questions, where you choose the right sequence of steps.
Independent practice material, not affiliated with Anthropic. Questions are original and not taken from any real exam. Always confirm exam details on Anthropic's official certification pages.