AIF-C01 Applications of Foundation Models Practice Question
A team is deploying a foundation model on Amazon Bedrock for a customer-facing assistant. They must reduce hallucinations and keep answers grounded in approved company content while controlling inference cost. Which TWO approaches should the team implement? (Choose two.)
⚠ Common exam trap
The trap here is treating model size or creativity settings as accuracy controls, when grounding requires retrieved source content plus verification of the response against it.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Configure Amazon Bedrock Guardrails contextual grounding and relevance checks to filter ungrounded responses.
Grounding answers in approved content through Bedrock knowledge bases supplies factual context, while Guardrails contextual grounding and relevance checks verify that generated responses are supported by that context. Together they reduce hallucination and keep outputs aligned with company material. Raising temperature, removing instructions, or simply choosing a larger model does not provide grounding and can increase cost or fabrication risk.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Remove all system instructions so the model relies solely on its pretrained knowledge.
Why it's wrong here
System instructions shape behavior, scope, and tone, and removing them leaves the model free to answer from potentially outdated or invented knowledge. Pretrained knowledge cannot reflect current approved company content and increases hallucination risk. Instructions also help the model use retrieved context correctly, so eliminating them is counterproductive.
- ✗
Switch to the largest available foundation model to guarantee factual accuracy.
Why it's wrong here
Model size does not guarantee truthfulness; larger models can still hallucinate confidently and typically cost more per token, conflicting with the cost-control requirement. Grounding and verification address accuracy more directly than scale. Selecting the largest model also reduces flexibility to right-size cost and latency for the workload.
- ✓
Configure Amazon Bedrock Guardrails contextual grounding and relevance checks to filter ungrounded responses.
Why this is correct
Guardrails contextual grounding evaluates whether a response is supported by the provided reference source and whether it is relevant to the user query, blocking or flagging outputs that fail thresholds. This adds a verification layer after generation, catching unsupported claims that retrieval alone may not prevent, and it strengthens grounding for customer-facing answers.
- ✗
Increase the model temperature to encourage more creative and varied responses.
Why it's wrong here
Higher temperature increases randomness in token selection, which raises the likelihood of fabricated or inconsistent statements. For a grounded customer assistant, this works against the goal of factual accuracy. Creativity is useful for brainstorming, not for policy-grounded question answering, so this setting would worsen hallucination rather than reduce it.
- ✓
Use Amazon Bedrock knowledge bases to retrieve relevant approved passages and include them in the prompt before generation.
Why this is correct
Retrieval-augmented generation via knowledge bases supplies the model with relevant, approved passages at query time, so answers are conditioned on source content rather than only parametric memory. This directly reduces hallucination and keeps responses aligned with company material. It also allows a smaller, cheaper model to perform well because the needed facts arrive in context instead of being memorized.
Go deeper
Related to this question
About these practice questions
One of 862 original AIF-C01 practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Amazon Web Services exam blueprint
This AIF-C01 practice question is part of Courseiva's free Amazon Web Services certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AIF-C01 exam.