mediumMultiple Choice
AIF-C01 Practice Question: A developer is using the Amazon Bedrock Converse…
A developer is using the Amazon Bedrock Converse API to build a multi-turn conversational AI. They need to send a user message along with system instructions and previous conversation history. How should they structure the API request to include both system prompt and message history?
⚠ Common exam trap
Watch out — candidates often confuse the Converse API's structure with simpler chat APIs (like OpenAI's) where system prompts are sometimes included in the messages array, leading them to incorrectly choose Option A or B.
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
✓
Include system prompt in the 'system' parameter and user/assistant messages in the 'messages' array
The Amazon Bedrock Converse API separates system prompts from the conversation history. System prompts are passed in the dedicated 'system' parameter, while user and assistant messages are placed in the 'messages' array. This design allows the model to distinguish persistent instructions from the ongoing dialogue, ensuring system-level guidance is not treated as part of the conversation context.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Include system prompt as a user message in the 'messages' array
Why it's wrong here
The Converse API has a dedicated 'system' parameter for system instructions; placing them in 'messages' as a user turn makes the model treat them as conversational input, so they compete with and can be overridden by later user turns. It is tempting because chat models do accept role-tagged user messages, which is the correct approach only for few-shot examples or user-supplied context.
- ✗
Include system prompt as an assistant message in the 'messages' array
Why it's wrong here
The Converse API exposes system instructions through the separate 'system' parameter; labelling them as an assistant message fabricates a prior model response, so the model treats the text as its own output rather than a governing instruction. It is tempting because assistant turns do steer subsequent replies, which works when seeding a conversation with an example response, not for policy.
- ✓
Include system prompt in the 'system' parameter and user/assistant messages in the 'messages' array
Why this is correct
The Converse API separates instructions from dialogue: the 'system' parameter carries the system prompt, while the 'messages' array holds alternating user and assistant turns. This structure preserves conversation history and satisfies the requirement to send both system instructions and prior context.
- ✗
Include system prompt in the 'inferenceConfig' parameter
Why it's wrong here
'inferenceConfig' carries generation parameters such as maxTokens, temperature, topP and stop sequences; it has no field for instructions, so the system prompt would be discarded entirely. It is tempting because inferenceConfig is where model behaviour is tuned, and temperature or stop sequences are the right lever when the goal is controlling output length or randomness rather than supplying instructions.
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 by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
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.