CCAO-F Domain 1 Study Guide: Prompting and Task Execution

Purpose of this study guide

This guide helps candidates prepare for Domain 1 of the Claude Certified Associate - Foundations exam. It translates the official blueprint into practical explanations, examples, decision rules, exercises, and original practice questions.

The official exam guide is the primary source. Anthropic describes this certification as an applied credential for professionals who use Claude to complete business and productivity tasks with minimal guidance. Typical candidates work in operations, marketing, project management, education, communications, consulting, or other knowledge-work roles.

This is not an API-development or enterprise-architecture certification. Candidates are not expected to design agentic systems or enterprise integrations. They are expected to use Claude effectively in professional workflows, recognize its limitations, and escalate more technical work to Claude Architects or Developers.

Supporting blog articles can deepen preparation, but they should be used after the official blueprint has established what must be learned. No blog, course, or question bank guarantees a passing result.

Domain 1 at a glance

Official domain title: Prompting and Task Execution

Official weight: 14% of the exam

Approximate item count: The exam contains 60 items. Fourteen percent corresponds to approximately 8 items. The actual number may vary because blueprint percentages are approximate.

Official objectives:

1. Create effective prompts for business and technical tasks. 2. Apply task decomposition techniques to structure complex requests. 3. Iterate prompts to improve output quality. 4. Adapt prompting strategies based on task type: analysis, research, drafting, and brainstorming.

What the domain is testing:

Can you turn a real work requirement into a clear, usable instruction for Claude, choose a sensible way to break down complex work, and improve the interaction when the first result is not good enough?

Exam context you should know

According to the official guide:

• •Exam code: CCAO-F.
• •Number of items: 60.
• •Item types: multiple-choice and multiple-response; each item states how many responses to select.
• •Time limit: 120 minutes.
• •Delivery: proctored online and/or at a test center, according to program policy.
• •Passing score: 720 on a scaled range of 100-1,000.
• •Exam fee: USD 99.
• •Credential validity: 12 months from the award date.
• •Result report: pass/fail, scaled score, and percent correct by domain.

The exam is criterion-referenced. You are measured against a fixed performance standard, not ranked against other candidates.

There are no mandatory prerequisites or required courses. Anthropic recommends regular hands-on professional use of Claude, foundational structured problem-solving and workflow skills, and a practical understanding of hallucinations, context limitations, and data sensitivity.

Always check the current Partner Academy, exam guide, registration page, and Pearson VUE confirmation because policies and details can change.

The core mindset: prompt for a work outcome

A prompt is not merely a question. In professional use, it is a work instruction. It should give Claude enough direction to produce an outcome that is relevant, usable, and appropriate for the situation.

Compare:

Weak: “Improve this report.”

Stronger: “Rewrite the attached weekly operations report for the regional director. Begin with the three issues requiring a decision, preserve all verified figures, explain operational impact in plain language, and limit the executive summary to 200 words. If a figure is unsupported by the report, flag it rather than correcting it by assumption.”

The stronger prompt clarifies:

• •what Claude must do.
• •who will read the result.
• •what information must be preserved.
• •what structure is expected.
• •how to handle uncertainty.

The objective is not to make every prompt long. The objective is to remove ambiguity that could materially affect the result.

A practical prompt structure: GOAL-C

Use GOAL-C as a study and workplace framework.

G - Goal

State the exact outcome.

Weak: “Analyze these survey responses.”

Better: “Identify the five recurring causes of employee dissatisfaction and recommend three actions HR can take within 90 days.”

O - Operating context

Provide the audience, purpose, relevant background, and authoritative material.

Example: “The findings will be presented to the HR leadership team. Use only the attached survey responses and the current employee policy.”

A - Actions

Describe the work Claude should perform. For a complex task, list meaningful stages.

Example: “Group similar comments, quantify each theme, identify representative quotations, compare themes across departments, and then recommend actions.”

L - Limits

State constraints, exclusions, tone, length, confidentiality, and missing-information behavior.

Example: “Do not identify individual employees. Do not infer demographic information. If a department has fewer than five responses, combine it into ‘Other’.”

C - Completion

Define the output and what finished means.

Example: “Return a one-page leadership brief containing a ranked theme table, three recommendations, and a short methodology note.”

GOAL-C is a preparation aid, not an official Anthropic exam acronym. Candidates should understand the underlying reasoning rather than memorize the letters alone.

Objective 1 - create effective prompts

An effective prompt commonly includes:

• •a direct task.
• •relevant context.
• •the intended audience.
• •necessary inputs or references.
• •clear constraints.
• •the desired output.
• •a rule for missing or uncertain information.

6.1 Be explicit about the action

Claude distinguishes between discussing work and doing work.

“Can you suggest improvements?” may produce recommendations.

“Rewrite the document using the improvements and then list the major changes” requests execution.

Use verbs that match the outcome:

• •analyze.
• •compare.
• •summarize.
• •classify.
• •research.
• •draft.
• •rewrite.
• •brainstorm.
• •prioritize.
• •create.
• •review.

6.2 Identify the audience and purpose

The same information should be communicated differently to a customer, senior leader, student, technical colleague, or general audience.

Instead of: “Explain the implementation.”

Use: “Explain the implementation to a nontechnical department head who must approve the budget. Focus on business benefits, timeline, major risks, and decisions required.”

6.3 Give relevant context

Useful context may include:

• •business objective.
• •organizational terminology.
• •current situation.
• •prior decisions.
• •approved reference material.
• •intended use of the output.

More context is not always better. Irrelevant history and duplicated information can distract from the task. Include what changes the answer.

6.4 Define constraints positively and specifically

Vague: “Do not make it too long or technical.”

Specific: “Use a maximum of five bullets. Keep each bullet under 25 words. Write for a nontechnical reader and define any unavoidable technical term.”

Negative instructions remain useful for real boundaries:

“Do not include customer names.”

Where possible, add the desired alternative:

“Use customer segments and aggregate counts instead of names.”

6.5 Specify missing-information behavior

Claude should not be encouraged to fill gaps with plausible-looking information.

Example: “Use only the supplied policy. If the policy does not answer a question, state ‘Not addressed in the supplied policy’ and list the information required.”

6.6 Separate instructions from source material

Use headings, tags, or clearly labeled sections:

TASK Summarize the complaint and recommend a routing category.

ROUTING POLICY [approved policy]

CUSTOMER COMPLAINT [customer text]

OUTPUT Category, urgency, evidence, and missing information.

The structure helps Claude distinguish the instruction from the content being analyzed.

6.7 Use examples when the pattern is difficult

Examples are helpful for tone, classification boundaries, formatting, and organization-specific conventions.

Select examples that are:

• •relevant.
• •correct.
• •varied.
• •consistent with the written instructions.

A useful set may include a normal case, an ambiguous case, and a missing-information case. Repeating many nearly identical examples adds less value than covering meaningful boundaries.

Objective 2 - apply task decomposition

Task decomposition means dividing complex work into manageable stages. It is useful when a request contains multiple objectives, depends on intermediate results, uses several source materials, or requires review before completion.

7.1 When decomposition is useful

Break down the task when:

• •the request combines research, analysis, and writing.
• •later work depends on earlier findings.
• •quality checks are important.
• •the user may need to approve direction before substantial work continues.
• •the output contains distinct deliverables.

7.2 A practical decomposition pattern

For a market briefing:

Step 1: Clarify the decision the briefing must support. Step 2: Research approved sources and record dates and links. Step 3: Separate verified facts from interpretation. Step 4: Identify three implications for the organization. Step 5: Draft the briefing for the named audience. Step 6: Check the draft against the length, evidence, and tone requirements.

7.3 Preserve dependencies

Do not ask Claude to recommend actions before it has analyzed the facts on which the recommendations depend. Sequence work logically.

Good sequence:

Extract facts -> compare evidence -> identify findings -> recommend action -> draft final output.

7.4 Do not over-decompose simple work

“Correct the spelling in this paragraph” does not require a six-stage workflow. Excessive steps add effort without improving the result.

Exam decision rule:

Use decomposition when it reduces complexity, preserves dependencies, creates useful review points, or improves accuracy. Use a direct prompt for simple, low-risk tasks.

7.5 Use checkpoints for consequential work

An Associate may ask Claude to analyze options and prepare a recommendation, while a person retains the final decision.

Example:

“First produce an outline and list any missing information. Wait for approval before drafting the final external communication.”

This creates a clear handoff and prevents unnecessary rework.

Objective 3 - iterate prompts to improve quality

The first output is a draft, not automatic proof that the prompt worked. Iteration should be purposeful: diagnose the failure, change the relevant instruction, and compare the result.

8.1 The REVIEW iteration loop

R - Result What did Claude produce?

E - Expected outcome What should the result have contained or avoided?

V - Variance Where does the output differ from the requirement?

I - Instruction change What specific prompt change addresses that difference?

E - Execute again Run the revised prompt on the same and additional representative inputs.

W - Weigh the evidence Did quality improve without causing a new problem?

REVIEW is a study framework, not an official exam acronym.

8.2 Diagnose before rewriting

Common problem -> likely prompt improvement

Too general -> add objective, audience, and evaluation criteria.

Wrong tone -> identify audience, relationship, purpose, and desired tone.

Missing sections -> provide an output structure or checklist.

Unsupported claims -> restrict sources and define missing-information behavior.

Inconsistent classification -> clarify label boundaries and add examples.

Too long -> add measurable length and priority constraints.

Claude suggests instead of acts -> use an explicit execution verb and completion condition.

8.3 Change one meaningful element at a time

If every part of a prompt changes, it is difficult to know what improved the result. For important recurring work, keep representative test cases and compare versions.

8.4 Do not ask Claude only to self-rate

Claude can help critique a draft, but its confidence is not proof of correctness. Improvement should be judged against the task requirements and, where factual claims matter, authoritative sources. Detailed validation belongs to Domain 2, but Domain 1 candidates should already understand that a polished answer may still be wrong.

8.5 Know when prompting is not the problem

If the task needs current information, use an appropriate research capability or source. If it needs a specialized product feature, choose that feature. If it contains sensitive data, follow organizational policy. If it requires an API or enterprise integration, escalate to the appropriate technical professional.

Objective 4 - adapt prompts to the task type

Different tasks require different instructions. The official blueprint specifically names analysis, research, drafting, and brainstorming.

9.1 Analysis prompts

Purpose: Interpret supplied information, compare options, identify patterns, or support a decision.

Include:

• •the decision or question.
• •criteria for comparison.
• •authoritative inputs.
• •required evidence.
• •treatment of uncertainty.
• •desired conclusion or recommendation format.

Example:

“Compare the three vendor proposals using implementation time, total three-year cost, security requirements, support coverage, and exit risk. Use only the attached proposals. Show the evidence for each rating, identify missing information, and provide a recommendation with two conditions that must be resolved before contracting.”

Common mistake: Asking Claude to “choose the best vendor” without defining best.

9.2 Research prompts

Purpose: Find, verify, and synthesize information.

Include:

• •the research question.
• •date or recency requirement.
• •acceptable or preferred sources.
• •geographic or industry scope.
• •verification expectations.
• •citation requirements.
• •how to handle disagreement between sources.

Example:

“Research current Indian higher-education guidance on responsible generative-AI use. Prioritize official government and university sources published or updated in 2026. Distinguish national requirements from institutional recommendations. Cite every material claim with a direct link and publication date. If reliable sources disagree, present both positions.”

Common mistake: Requesting “latest information” without authorizing research or specifying sources and date.

9.3 Drafting prompts

Purpose: Create a document, message, plan, summary, or other communication.

Include:

• •format.
• •audience.
• •communication objective.
• •essential facts.
• •tone and voice.
• •length.
• •required call to action.
• •information that must not be invented.

Example:

“Draft a 180-220 word email from the operations director to department heads announcing the revised travel-approval process. Use a calm, practical tone. Explain what changes, when it starts, and the three actions recipients must take. Use only the supplied policy. End with the support contact. Do not add penalties or exceptions not stated in the policy.”

Common mistake: Specifying only “professional tone” without audience, purpose, or required facts.

9.4 Brainstorming prompts

Purpose: Generate diverse possibilities before evaluation.

Include:

• •the challenge or opportunity.
• •target user.
• •constraints.
• •desired number and diversity of ideas.
• •categories or dimensions to explore.
• •whether evaluation should happen now or later.

Example:

“Generate 15 ways a university can help first-year students use Claude responsibly. Include five classroom ideas, five student-support ideas, and five faculty-development ideas. At least five ideas must require no new software budget. Do not rank the ideas yet; prioritize variety in this first round.”

Common mistake: Combining idea generation and strict evaluation too early, which can reduce variety.

9.5 Choosing the correct strategy

Analysis -> define criteria and evidence. Research -> define sources, recency, verification, and citations. Drafting -> define audience, purpose, facts, format, and voice. Brainstorming -> define variety, quantity, categories, and constraints.

Some workflows combine task types. Decompose them:

Research -> analyze -> brainstorm options -> select with human judgment -> draft the final communication.

Common domain 1 exam traps

1. A role without a task

“Act as an expert consultant” does not define the required outcome.

2. Vague success words

“Excellent,” “professional,” “detailed,” and “brief” need observable meaning.

3. Missing audience

Claude cannot reliably choose language, depth, and tone without knowing who will use the output.

4. Too much irrelevant context

Extra information can increase distraction rather than improve the answer.

5. Conflicting constraints

“Comprehensive analysis of every issue in 100 words” needs prioritization or separate outputs.

6. Examples that contradict instructions

Claude may follow the demonstrated pattern, so examples must be correct.

7. No rule for missing facts

Without guidance, absent information may be handled inconsistently.

8. Treating every task as one step

Complex research-to-recommendation work benefits from decomposition.

9. Over-decomposing a simple request

More steps are not automatically better.

10. Repeating “be accurate” instead of fixing the prompt

Add sources, criteria, examples, and missing-information behavior.

11. Using the same prompt style for every task

Research, analysis, drafting, and brainstorming have different needs.

12. Changing the whole prompt after one weak result

Diagnose the failure and make a targeted revision.

13. Expecting prompting to solve a product, policy, or technical problem

Use the correct Claude feature, follow policy, or escalate when necessary.

14. Assuming a polished response is correct

Quality of presentation is not evidence of factual accuracy.

Rapid scenario-answering method

When answering a Domain 1 scenario, ask:

1. What work outcome is required? 2. What information must Claude use? 3. Who is the audience? 4. Is the task simple or should it be decomposed? 5. Which task type is primary: analysis, research, drafting, or brainstorming? 6. What constraints or missing-information rules matter? 7. What exactly was wrong with the existing output? 8. Which option fixes that problem most directly?

Prefer the answer that is clear, relevant, proportionate, and testable. Avoid an option that adds impressive wording but does not address the stated failure.

For multiple-response items, read the stated number of selections carefully. Judge each option independently before choosing the required combination.

Original practice questions

These questions were created for study and are not from the live exam.

QUESTION 1 - MULTIPLE CHOICE

An associate asks Claude, “Write a project update,” and receives a generic response. Which revised prompt is most effective?

A. “Act as the world’s best project manager and make it impressive.” B. “Write a project update.” C. “Draft a 150-word update for the steering committee. Summarize completed milestones, current risks, decisions required this week, and the next milestone. Use the attached status notes only and flag missing dates.” D. “Think harder and provide more details.”

Correct answer: C

Rationale: It identifies audience, format, required content, source boundary, length, and missing-information behavior.

QUESTION 2 - MULTIPLE RESPONSE - SELECT TWO

A recurring classification prompt produces inconsistent results for borderline cases. Which two improvements are most appropriate?

A. Define the boundaries between categories. B. Add representative examples of borderline cases. C. Ask Claude to use a more confident tone. D. Add unrelated background about the organization.

Correct answers: A and B

Rationale: Clear criteria and boundary examples directly address inconsistency. Tone and irrelevant context do not.

QUESTION 3 - MULTIPLE CHOICE

A manager needs a recommendation based on four reports. What is the best first step?

A. Ask Claude to recommend an option immediately. B. Ask Claude to extract and compare the relevant evidence using stated decision criteria before recommending. C. Ask Claude to brainstorm unrelated alternatives. D. Ask Claude to rewrite all reports in a more formal tone.

Correct answer: B

Rationale: The recommendation depends on evidence. Decomposition preserves that dependency.

QUESTION 4 - MULTIPLE CHOICE

Claude’s email draft is too technical for customers. Which prompt change best addresses the problem?

A. “Make it better.” B. “Write as an expert engineer.” C. “Rewrite for customers with no technical background. Use plain language, define the one unavoidable technical term, and focus on what changes for the customer.” D. “Add more technical details for accuracy.”

Correct answer: C

Rationale: It defines the audience and gives observable language and content instructions.

QUESTION 5 - MULTIPLE RESPONSE - SELECT THREE

Which three elements are especially important in a research prompt?

A. Recency or date requirements B. Preferred or authoritative sources C. Verification and citation expectations D. A fictional expert persona E. A guarantee that every source will agree

Correct answers: A, B, and C

Rationale: Research quality depends on scope, sources, and verification. A persona does not replace these controls, and sources may legitimately disagree.

QUESTION 6 - MULTIPLE CHOICE

A brainstorming session produces only minor variations of one idea. What is the best improvement?

A. Ask for more detail about the same idea. B. Define several categories to explore and request diverse ideas before ranking them. C. Ask Claude to select the safest idea immediately. D. Reduce the number of ideas to one.

Correct answer: B

Rationale: Category coverage and delayed evaluation encourage variety.

QUESTION 7 - MULTIPLE CHOICE

A policy summary includes a deadline not found in the supplied document. What Domain 1 prompt improvement would most directly reduce this risk?

A. Ask for a more formal summary. B. State that Claude must use only the supplied policy and label information not addressed in it. C. Ask Claude to report high confidence. D. Increase the word limit.

Correct answer: B

Rationale: The instruction establishes an authoritative source and honest missing-information behavior. Domain 2 covers the additional validation step.

QUESTION 8 - MULTIPLE RESPONSE - SELECT TWO

Claude produces a long answer when the user needs a concise leadership brief. Which two changes best improve the prompt?

A. Specify the leadership audience and decisions required. B. Set an observable format and length limit. C. Add the entire conversation history. D. Ask Claude to be generally more intelligent.

Correct answers: A and B

Rationale: Audience, purpose, format, and measurable length directly address the failure.

Seven-day domain 1 study plan

Use this plan after reading the official exam guide and Domain 1 objectives. Supporting blog articles should reinforce each day’s topic, not replace hands-on practice.

DAY 1 - Understand the candidate level

Read Sections 1-7 of the official exam guide. Write a one-page summary of what an Associate should do and what should be escalated to Architects or Developers.

Blog support: Use the role-fit and certification-value articles to understand whether the credential matches your work.

DAY 2 - Build clear business prompts

Rewrite ten vague workplace prompts using GOAL-C. Include operations, marketing, education, communications, and project-management examples.

Blog support: Use the Domain 1 article as the main reference for prompt structure and common traps.

DAY 3 - Practice task decomposition

Take five complex requests and split them into logical stages with dependencies and review points. Take five simple requests and explain why they should remain direct.

Blog support: Use the first-attempt anti-pattern article to recognize unnecessary complexity and weak control choices.

DAY 4 - Adapt by task type

Write one strong prompt each for analysis, research, drafting, and brainstorming. Then convert a single business challenge through all four stages.

Blog support: Use supporting prompt and workflow articles for additional examples, while checking them against the official objectives.

DAY 5 - Iterate systematically

Run three prompts in Claude. Diagnose each weak result using REVIEW, revise one meaningful element, and compare the outputs.

Blog support: Use the troubleshooting and practice-question articles to build an error log.

DAY 6 - Timed scenario practice

Answer the eight questions in this guide and a fresh set of original questions. Practice both multiple-choice and multiple-response instructions.

Blog support: Use the free practice-question article, but do not treat practice scores as a guaranteed exam prediction.

DAY 7 - Consolidate and prepare

Review mistakes, create a one-page Domain 1 checklist, and revisit only the official documentation related to your gaps.

Blog support: Use the exam-day strategy article for logistics and time management. Confirm all operational details in the latest official materials.

Final readiness checklist

[ ] I can explain the four official Domain 1 objectives. [ ] I can create a clear prompt for a real business task. [ ] I can define audience, context, constraints, and output. [ ] I know when examples will improve a prompt. [ ] I can represent missing information without encouraging invention. [ ] I can decompose complex work while keeping simple work direct. [ ] I can diagnose a weak output and make a targeted prompt revision. [ ] I can adapt prompts for analysis, research, drafting, and brainstorming. [ ] I understand that polished wording does not prove factual correctness. [ ] I know when to use another Claude feature, follow policy, seek human review, or escalate technical work. [ ] I can answer both multiple-choice and multiple-response practice items carefully.

How to use the related blog library

The blog library should be organized around candidate questions:

• •“What does Domain 1 cover?” -> this Domain 1 study guide.
• •“How should I study?” -> seven-day study plan and career-switch preparation guide.
• •“What mistakes should I avoid?” -> first-attempt anti-pattern guide.
• •“Am I ready?” -> practice questions and readiness checklist.
• •“Is this certification right for me?” -> role-fit and honest-value assessment.
• •“What should I expect on exam day?” -> exam-day strategy guide.

Editorial rule for every blog:

1. Use the official Exam Guide Version 1.0 as the source for exam facts. 2. Clearly label original practice questions as non-live content. 3. Do not promise a pass, job, promotion, or financial return. 4. Link recommendations back to the relevant official objective. 5. Recheck time-sensitive exam policies before publishing updates.

Sources

PRIMARY EXAM SOURCE

Claude Certified Associate - Foundations Exam Guide Version 1.0 | Effective July 2026 | Exam code CCAO-F Attached reference file: ClaudeCertifiedAssociateFoundationsExamGuide.pdf

OFFICIAL ANTHROPIC LEARNING AND PRODUCT SOURCES

Anthropic learning resources: https://www.anthropic.com/learn

Anthropic Academy - Build with Claude: https://www.anthropic.com/learn/build-with-claude

Prompting best practices for current Claude models: https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/claude-prompting-best-practices

Prompt engineering overview: https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/overview

Define success criteria and build evaluations: https://platform.claude.com/docs/en/test-and-evaluate/develop-tests

Note: Official product documentation may contain developer-oriented material beyond the Associate exam scope. Study only the portions relevant to the Domain 1 objectives and professional Claude usage.

End of domain 1 study guide