The most effective AIP-C01 study plan runs 6 to 8 weeks and front-loads the two heaviest domains. Spend your first two weeks building a real RAG pipeline in Amazon Bedrock (Domain 1, 31%), weeks three and four on Bedrock Agents and integration (Domain 2, 26%), weeks five and six on safety, optimization, and troubleshooting (Domains 3–5), and the final two weeks on timed practice exams until you consistently hit 85%.
This is a hands-on, production-focused exam — the AWS Certified Generative AI Developer – Professional (AIP-C01) rewards people who build with Bedrock, not people who memorize. The schedule below maps each week directly to the five official exam domains so your study time tracks the scored weighting instead of treating every topic equally.
The 8-week AIP-C01 study plan
Eight weeks is the sweet spot for most candidates — enough time to build real systems without losing momentum. The plan is structured so that every week produces something you've actually deployed, because that's how the questions are framed.
| Weeks | Focus | Domains | Build target |
|---|---|---|---|
| 1–2 | Foundation models & RAG | Domain 1 (31%) | A working RAG pipeline in Amazon Bedrock with a Knowledge Base |
| 3–4 | Agents & integration | Domain 2 (26%) | Bedrock Agents with action groups + a multi-step Prompt Flow |
| 5–6 | Safety, optimization, troubleshooting | Domains 3 (20%), 4 (12%), 5 (11%) | Guardrails config, cost-optimized inference, a broken RAG pipeline you debug |
| 7–8 | Timed practice & gap closing | All five | 85%+ on full-length practice exams, twice |
Weeks 1–2: Foundation Model Integration & RAG (Domain 1)
Domain 1 is the single largest at 31% of scored content, so it gets the most calendar time. Don't read about RAG — build one. Stand up an Amazon Bedrock Knowledge Base backed by a vector store, ingest a document set, and wire up retrieval-augmented generation end to end. Along the way you'll naturally cover foundation-model selection, the InvokeModel and Converse APIs, embeddings, chunking strategy, and the data-management and compliance concerns (PII handling, data residency) that this domain bundles in.
By the end of week two you should be able to explain why you chose a given model, chunk size, and vector store — because the exam asks you to reason about those tradeoffs, not recite them.
Weeks 3–4: Implementation and Integration (Domain 2)
Domain 2 is the second-heaviest at 26%, and together with Domain 1 it makes up 57% of the exam — get these two right and you've covered most of the test. This is the agentic core: build Bedrock Agents with action groups that call your own Lambda functions, then chain multi-step logic with Prompt Flows. If you've worked through the exam guide, you'll recognize that Amazon Bedrock AgentCore and agentic workflows are exactly what the refreshed exam emphasizes — so the hours you spend here are the highest-leverage of the whole plan.
Wire at least one agent to a real downstream system (an API, a database lookup). The questions present code snippets and ask you to reason about API usage, so reading your own working integration code is the best possible preparation.
Weeks 5–6: Safety, Optimization & Troubleshooting (Domains 3–5)
These three domains total 43% combined and are the ones candidates most often underestimate. Domain 3 (AI Safety, Security, and Governance, 20%) is substantial on its own: configure Bedrock Guardrails, set up content filtering and denied topics, and understand IAM-scoped access to models. Domain 4 (Operational Efficiency, 12%) is concrete and high-yield — provisioned vs on-demand throughput, prompt caching, model distillation, and cost-optimization patterns. Domain 5 (Testing, Validation, and Troubleshooting, 11%) is the lifecycle's tail: deliberately break your RAG pipeline (bad chunking, wrong embeddings model, a misconfigured retrieval filter) and practice diagnosing it. The fastest way to lock in Domain 5 is to have actually fixed the failure modes it describes.
Weeks 7–8: Timed practice exams
The final two weeks are pure exam simulation. Take full-length, timed practice exams under real conditions, then review every wrong answer until you understand the underlying tradeoff — not just the right letter. Rotate through original AIP-C01 practice questions so you're testing reasoning, not memorizing a dump. The bar to book the real exam: 85% or higher, consistently, across all five domains — confirm you're hitting it on the safety, optimization, and troubleshooting sections specifically, since those are the easy ones to neglect.
The 4-week fast track
If you already build with Amazon Bedrock daily and know AWS core services cold, you can compress the plan to four weeks:
- Week 1: Refresh Domain 1 — re-build a RAG pipeline only if yours isn't current. Skim the FM-integration and compliance topics.
- Week 2: Domain 2 — Bedrock Agents, action groups, Prompt Flows. This is where even experienced builders find gaps.
- Week 3: Domains 3–5 — Guardrails, cost optimization, and troubleshooting in one focused sprint.
- Week 4: Timed practice exams until you hit 85%+ twice.
I'd be honest about who this is for, though. The compressed plan works when the foundational week is genuinely a refresher. If you find yourself learning RAG architecture for the first time in week one, you're on the 8-week plan whether you scheduled it or not — and the architectural reasoning you build over the longer runway is exactly what separates a pass from a near-miss on the harder questions.
How to know you're ready
You're ready to book when three things are true: you can build a RAG pipeline and a Bedrock Agent from scratch without reference, you can read a code-snippet question and reason about the API behavior, and you're scoring 85%+ on timed full-length practice exams across every domain. If your safety or troubleshooting scores lag your RAG scores, you're not done — those domains are 31% of the exam combined.
For the exact exam mechanics behind this plan — the passing score, cost, and prerequisites — see the dedicated guides. And because this exam rewards builders, the highest-leverage thing you can do is spend the hands-on weeks in a real environment. Practice with CloudaQube's hands-on AI labs to build the RAG pipelines, agents, and Guardrails configs this plan is built around — the same way the exam tests them.
Sources: official AWS Certified Generative AI Developer – Professional exam guide for domain weights, plus current study-timeline data aggregated from KodeKloud and Whizlabs. Confirm domain weights against the official AWS exam guide before you start.