CLAUDE CERTIFIED ARCHITECT – FOUNDATIONS · EXAM PREP

Pass the
Claude Architect
exam by building
what it tests.

Self-paced prep material across all five exam domains and all six scenarios: video lessons, slides, quizzes, hands-on labs and timed mock exams, on a platform that shows when you are ready to book.

Your readiness
SAMPLE VIEW
MOCK 02 · SCALED SCORE
684
/ 1,000
36 points to pass
100
720 TO PASS
1,000
Agentic Architecture & Orchestration
78%
Claude Code Configuration & Workflows
84%
Prompt Engineering & Structured Output
71%
Tool Design & MCP Integration
52%
Context Management & Reliability
66%
Fix Tool Design first, then sit another mock.
60
questions
Multiple choice and multiple response
120
minutes
Proctored and closed book
720
to pass
On a scale of 100 to 1,000
$125
exam fee
Paid again on every retake
WHY PREPARE

This exam tests judgment,
not memory.

You get 4 production scenarios, picked at random from a set of 6, and 60 questions on what you would actually do in each. It is proctored and closed book. Anthropic expects candidates to have around 6 months of hands-on building with Claude.

Built on real problems.
Questions sit inside production situations, from support agents to CI pipelines.
No official practice exam.

Anthropic retired its practice exam when the program moved to Pearson. Our timed mocks give you that rehearsal.

Failing costs more than the fee.

Every retake is full price, and the wait grows each time.

THE COST OF SITTING IT UNPREPARED
Three failures: 134 days of waiting and four full exam fees.
$500
at list price
14 days
After 1st fail
30 days
After 2nd fail
90 days
After 3rd fail
WHAT YOU GET

Everything in
the prep material

THE PART QUESTION BANKS SKIP
Hands-on labs

Guided builds, so you have already made the kind of decisions the exam asks about.

EXAMPLE
A multi-tool agent with escalation
EXAMPLE
A Claude Code team workflow
EXAMPLE
A structured data extraction pipeline
EXAMPLE
A multi-agent research pipeline
Timed mock exams

Full-length practice in the exam's format, scored by domain. The rehearsal the official program no longer offers.

QUESTION 34 OF 60
51:08
Scenario 3 of 4: Multi-agent research system
Video lessons
Lessons organized around the exam's five domains.
Slides
Slides to study and revise from.
Quizzes
Quick checks as you go, so gaps show up early.
Readiness view

Your score in each domain, so you know what to fix before paying for the exam.

Today you can buy questions, or a live bootcamp.
This teaches you, then tests you.

Question banks
Live bootcamps
Our prep material
Teaches the material
Hands-on labs
Timed mock exams
Readiness by domain
Some
Self-paced
Team progress view
Some
Get early access
WHAT THE EXAM COVERS

All five domains,
as the exam weights them

27%
Agentic Architecture & Orchestration
Agentic loops, coordinator and subagent systems, multi-step workflows, SDK hooks, task decomposition, session state
20%
Claude Code Configuration & Workflows
CLAUDE.md hierarchy, commands and skills, path rules, plan mode, iterative refinement, CI/CD
20%
Prompt Engineering & Structured Output
Explicit criteria, few-shot prompting, JSON schemas through tool use, validation and retries, batch processing, multi-pass review
18%
Tool Design & MCP Integration
Clear tool interfaces, structured errors, distributing tools across agents, MCP servers, built-in tools
15%
Context Management & Reliability
Long-conversation context, escalation, error propagation, large codebases, human review, information provenance
THE SIX SCENARIOS

Prepare for all six.
The exam picks four.

The draw is random, so you need to be ready for any of them. Here is one possible draw.

DRAWN
Customer support resolution agent
NOT THIS TIME
Code generation with Claude Code
DRAWN
Multi-agent research system
DRAWN
Developer productivity with Claude
NOT THIS TIME
Claude Code for continuous integration
DRAWN
Structured data extraction
THE LEARNING EXPERIENCE

See how
you will learn

Progress
DOMAIN 1 · 27%
Agentic loops
Task decomposition
Coordinators and subagents
SDK hooks
Session state
Multi-step workflows
Coordinators and subagents
SAMPLE VIEW
QUICK CHECK
A subagent returns malformed output on its third retry. What should the coordinator do?
Retry again with the same prompt
Return a structured error and escalate
Skip the subagent and continue silently
01
Learn and check.
A sample lesson, with a quick check below the video.
Your five domains
SAMPLE VIEW
78%
Agentic
84%
Claude Code
71%
Prompts
52%
Tools
66%
Context
02
Know where you stand.
Your readiness across the five domains.
HOW IT WORKS

From payment to
your first lesson

1
Buy
the prep material.
2
Check your email

for a one-time access code.

3
Create your account
on the Product Faculty learning platform.
4
Enter your code.
Your prep material opens.

When your mocks say you are ready, book the exam through Anthropic's Partner Academy.

Your one-time access code
From Product Faculty
Thanks for your purchase. Use this code once to open all of the prep material.
PFCA-7K2M-9QXD
Create your account
Code accepted. All five domains are open.
FOR TEAMS

Certifying a
whole practice?

Certified people count toward your standing in the Claude Partner Network. See who is ready, in which domain, before you spend an exam fee on them.

Access codes for your whole team
Progress and readiness for each person
Get early access
Cloud practice · 24 seats
SAMPLE VIEW
9
Ready to book
11
Still in prep
4
Not started
Person
Agent
Code
Prompt
Tools
Context
Status
A. Rahman
84
81
78
76
80
Book now
J. Lindqvist
79
88
75
77
76
Book now
P. Nair
72
80
66
52
70
Fix Tools
M. Okafor
58
74
69
63
55
Fix Agent
S. Tanaka
66
70
64
62
63
In prep
36,000+
consultants certified
1,300+
organizations
PUBLIC PLEDGES · SOURCE: ANTHROPIC, 23 JUL 2026
50,000
Accenture
30,000
PwC
15,000
Deloitte
15,000
KPMG
WHO BUILT IT

Built by practitioners
who ship with Claude

4,000+
product managers trained by Product Faculty, with instructors from OpenAI, Google and Uber.
QUESTIONS

Questions,
answered straight

Something else? Write to support@productfaculty.com.

No. This is preparation. Anthropic runs the certification and Pearson delivers the exam.

No. Exam content is confidential. Everything we teach is written from Anthropic's public exam guide, and our practice questions are our own.

No. The exam is $125, paid to Anthropic when you register. Some partner tiers get a discount.

Not at the moment. You need a job at a Claude Partner Network organization.

A digital badge from Anthropic through Credly, valid for 12 months. You renew it with a free assessment.

You book your own slot. Sign in to Anthropic's partner site with your company email, choose the exam and pay there. Anthropic then emails you a Pearson VUE link to book a time, online with a proctor or at a test center.

120 minutes, and about 135 in the seat. You answer 60 multiple choice and multiple response questions, and you need 720 on a scale of 100 to 1,000 to pass.

Four scenarios, drawn at random from a set of six, and questions across five domains: agentic architecture, Claude Code, prompt engineering, tool design and MCP, and context management.

You can retake it, at the full fee each time. You wait 14 days after the first fail, 30 after the second and 90 after the third, with up to 4 attempts in 12 months.

Know you are ready before you book.
Claude and Anthropic are trademarks of Anthropic, PBC. This material is an independent preparation resource and is not written or endorsed by Anthropic or Pearson.
CLAUDE CERTIFIED ARCHITECT – FOUNDATIONS · EXAM PREP
Pass the
Claude Architect
exam by building what it tests.

Self-paced prep material across all five exam domains and all six scenarios: video lessons, slides, quizzes, hands-on labs and timed mock exams, on a platform that shows when you are ready to book.

Your readiness
SAMPLE VIEW
MOCK 02 · SCALED SCORE
684
/ 1,000
36 points to pass
100
720 TO PASS
1,000
Agentic Architecture & Orchestration
78%
Claude Code Configuration & Workflows
84%
Prompt Engineering & Structured Output
71%
Tool Design & MCP Integration
52%
Context Management & Reliability
66%
Fix Tool Design first, then sit another mock.
60
questions
Multiple choice and multiple response
120
minutes
Proctored and closed book
720
to pass
On a scale of 100 to 1,000
$125
exam fee
Paid again on every retake
WHY PREPARE

This exam tests judgment,
not memory.

You get 4 production scenarios, picked at random from a set of 6, and 60 questions on what you would actually do in each. It is proctored and closed book. Anthropic expects candidates to have around 6 months of hands-on building with Claude.

Built on real problems.
Questions sit inside production situations, from support agents to CI pipelines.
No official practice exam.

Anthropic retired its practice exam when the program moved to Pearson. Our timed mocks give you that rehearsal.

Failing costs more than the fee.

Every retake is full price, and the wait grows each time.

THE COST OF SITTING IT UNPREPARED
Three failures: 134 days of waiting and four full exam fees.
$500
at list price
14 days
After 1st fail
30 days
After 2nd fail
90 days
After 3rd fail
WHAT YOU GET

Everything in
the prep material

THE PART QUESTION BANKS SKIP
Hands-on labs

Guided builds, so you have already made the kind of decisions the exam asks about.

EXAMPLE
A multi-tool agent with escalation
EXAMPLE
A Claude Code team workflow
EXAMPLE
A structured data extraction pipeline
EXAMPLE
A multi-agent research pipeline
Timed mock exams

Full-length practice in the exam's format, scored by domain. The rehearsal the official program no longer offers.

QUESTION 34 OF 60
51:08
Scenario 3 of 4: Multi-agent research system
Video lessons
Lessons organized around the exam's five domains.
Slides
Slides to study and revise from.
Quizzes
Quick checks as you go, so gaps show up early.
Readiness view

Your score in each domain, so you know what to fix before paying for the exam.

Today you can buy questions, or a live bootcamp.
This teaches you, then tests you.

Question banks
Live bootcamps
Our prep material
Teaches the material
Hands-on labs
Timed mock exams
Readiness by domain
Some
Self-paced
Team progress view
Some
Get early access
WHAT THE EXAM COVERS

All five domains,
as the exam weights them

27%
Agentic Architecture & Orchestration
Agentic loops, coordinator and subagent systems, multi-step workflows, SDK hooks, task decomposition, session state
20%
Claude Code Configuration & Workflows
CLAUDE.md hierarchy, commands and skills, path rules, plan mode, iterative refinement, CI/CD
20%
Prompt Engineering & Structured Output
Explicit criteria, few-shot prompting, JSON schemas through tool use, validation and retries, batch processing, multi-pass review
18%
Tool Design & MCP Integration
Clear tool interfaces, structured errors, distributing tools across agents, MCP servers, built-in tools
15%
Context Management & Reliability
Long-conversation context, escalation, error propagation, large codebases, human review, information provenance
THE SIX SCENARIOS

Prepare for all six.
The exam picks four.

The draw is random, so you need to be ready for any of them. Here is one possible draw.

DRAWN
Customer support resolution agent
NOT THIS TIME
Code generation with Claude Code
DRAWN
Multi-agent research system
DRAWN
Developer productivity with Claude
NOT THIS TIME
Claude Code for continuous integration
DRAWN
Structured data extraction
THE LEARNING EXPERIENCE

See how
you will learn

Progress
DOMAIN 1 · 27%
Agentic loops
Task decomposition
Coordinators and subagents
SDK hooks
Session state
Multi-step workflows
Coordinators and subagents
SAMPLE VIEW
QUICK CHECK
A subagent returns malformed output on its third retry. What should the coordinator do?
Retry again with the same prompt
Return a structured error and escalate
Skip the subagent and continue silently
01
Learn and check.
A sample lesson, with a quick check below the video.
Your five domains
SAMPLE VIEW
78%
Agentic
84%
Claude Code
71%
Prompts
52%
Tools
66%
Context
02
Know where you stand.
Your readiness across the five domains.
HOW IT WORKS

From payment to
your first lesson

1
Buy
the prep material.
2
Check your email

for a one-time access code.

3
Create your account
on the Product Faculty learning platform.
4
Enter your code.
Your prep material opens.

When your mocks say you are ready, book the exam through Anthropic's Partner Academy.

Your one-time access code
From Product Faculty
Thanks for your purchase. Use this code once to open all of the prep material.
PFCA-7K2M-9QXD
Create your account
Code accepted. All five domains are open.
FOR TEAMS

Certifying a
whole practice?

Certified people count toward your standing in the Claude Partner Network. See who is ready, in which domain, before you spend an exam fee on them.

Access codes for your whole team
Progress and readiness for each person
Get early access
Cloud practice · 24 seats
SAMPLE VIEW
9
Ready to book
11
Still in prep
4
Not started
Person
Agent
Code
Prompt
Tools
Context
Status
A. Rahman
84
81
78
76
80
Book now
J. Lindqvist
79
88
75
77
76
Book now
P. Nair
72
80
66
52
70
Fix Tools
M. Okafor
58
74
69
63
55
Fix Agent
S. Tanaka
66
70
64
62
63
In prep
36,000+
consultants certified
1,300+
organizations
PUBLIC PLEDGES · SOURCE: ANTHROPIC, 23 JUL 2026
50,000
Accenture
30,000
PwC
15,000
Deloitte
15,000
KPMG
WHO BUILT IT

Built by practitioners
who ship with Claude

4,000+
product managers trained by Product Faculty, with instructors from OpenAI, Google and Uber.
QUESTIONS

Questions,
answered straight

Something else? Write to support@productfaculty.com.

No. This is preparation. Anthropic runs the certification and Pearson delivers the exam.

No. Exam content is confidential. Everything we teach is written from Anthropic's public exam guide, and our practice questions are our own.

No. The exam is $125, paid to Anthropic when you register. Some partner tiers get a discount.

Not at the moment. You need a job at a Claude Partner Network organization.

A digital badge from Anthropic through Credly, valid for 12 months. You renew it with a free assessment.

You book your own slot. Sign in to Anthropic's partner site with your company email, choose the exam and pay there. Anthropic then emails you a Pearson VUE link to book a time, online with a proctor or at a test center.

120 minutes, and about 135 in the seat. You answer 60 multiple choice and multiple response questions, and you need 720 on a scale of 100 to 1,000 to pass.

Four scenarios, drawn at random from a set of six, and questions across five domains: agentic architecture, Claude Code, prompt engineering, tool design and MCP, and context management.

You can retake it, at the full fee each time. You wait 14 days after the first fail, 30 after the second and 90 after the third, with up to 4 attempts in 12 months.

Know you are ready before you book.
Claude and Anthropic are trademarks of Anthropic, PBC. This material is an independent preparation resource and is not written or endorsed by Anthropic or Pearson.

Get early access

Tell us where to reach you and we will email you when it opens.

No payment today.