AI-102 Implement an agentic solution — Timed 10 Questions

Question 1

You are using Microsoft Copilot Studio to create an agent that handles product returns. The agent needs to collect the user's order number and reason for return. Which TWO features should you use to capture this information?

10 minute time limit — choose an answer to begin.

Accepted Answer

Use variables to store the order number and reason. Options C and D are correct. C: Variables are used in Copilot Studio to store user-provided information like order numbers and reasons during the conversation. D: Custom entities allow the agent to extract specific data from user input, such as order numbers and return reasons. Option A is incorrect because Power Automate flows are for automation after data capture, not for initial input parsing. Option B is incorrect because trigger phrases define when a topic is activated, not capture data. Option E is incorrect because topics manage conversation flows, not data storage.

Answer

Use Power Automate flows to parse the user's input

Answer

Define trigger phrases for each topic

Answer

Create separate topics for each type of return

Question 2

Refer to the exhibit. You deploy this ARM template to create an agent in Microsoft Foundry. The agent uses an Azure Function as an action. The deployment succeeds but when the agent calls the function, it gets a 401 Unauthorized error. What is the most likely cause?

Accepted Answer

The listKeys function returns the host key, but the function requires a function-level key.. When an Azure Function uses function-level authorization, each function requires its own specific key. The `listKeys` function returns the host key by default, which does not grant access to a specific function, resulting in a 401 Unauthorized error. Options A and B are incorrect because the authentication type (ApiKey vs. ManagedIdentity) and the model provider (AzureOpenAI vs. AzureAI) do not cause a 401 when calling a function. Option C is incorrect because the URL parameter is not missing the function name; the agent's endpoint is correctly constructed.

Answer

The authentication type should be ManagedIdentity instead of ApiKey.

Answer

The model provider is AzureOpenAI but should be AzureAI.

Answer

The URL parameter is missing the function name in the path.

Question 3

You are building an agent using Azure AI Agent Service that must execute code in a sandbox environment. The code should be able to install Python packages. Which action type should you use?

Accepted Answer

code_interpreter. (code_interpreter) is correct because it provides a sandboxed Python environment that allows installing packages and executing code securely. Option A (function) is incorrect because it is used for calling custom functions or APIs, not for running arbitrary code in a sandbox. Option C (openApi) is incorrect because it is for defining OpenAPI specifications for API endpoints, not for code execution. Option D (httpRequest) is incorrect because it is for making HTTP requests to external services, not for sandboxed code execution.

Answer

function

Answer

openApi

Answer

httpRequest

Question 4

Refer to the exhibit. You are sending a request to an agent using the Chat Completions API. The agent should decide whether to call the 'get_weather' function or not. What is missing from the request to enable the agent to properly use the function?

Accepted Answer

The 'messages' array should include a system prompt. A system prompt is missing. In the Chat Completions API, a system message sets the behavior and context for the assistant. Without it, the agent may not understand when to call the 'get_weather' function. The system prompt should instruct the assistant to use the function when appropriate.

Answer

The 'tool_choice' should be 'required'

Answer

The 'model' should be 'gpt-4o-mini'

Answer

The 'strict' parameter in the function definition

Question 5

Refer to the exhibit. You have created an assistant with the above configuration. When you send a message 'What is the weather in Seattle?', the assistant responds without calling the function. What is the most likely cause?

Accepted Answer

The function is missing the 'description' property. The function definition lacks a 'description' property. In the Assistants API, the 'description' field is critical for the model to understand when and why to invoke a function. Without it, the model may not recognize that the function is relevant to the user's query about weather, causing it to respond directly instead of calling the function.

Answer

The 'instructions' are not being followed

Answer

The 'strict' parameter is set to true

Answer

The 'tool_resources' code_interpreter has empty file_ids

Question 6

You are building an agentic solution using Microsoft Semantic Kernel. The agent must autonomously decide when to call an external API to fetch real-time data. You want to minimize token usage and avoid unnecessary API calls. Which planner configuration should you use?

Accepted Answer

Use an AutoInvoke kernel with a greedy action planner. An AutoInvoke kernel with a greedy action planner enables the agent to make step-by-step decisions about which function to invoke next based on the immediate context. This minimizes token consumption by forgoing exhaustive plan generation and only triggering the external API when the current step requires it. In contrast, SequentialPlanner generates a complete plan upfront, potentially including unnecessary API calls, and manual or parallel planners either lack autonomy or increase token overhead.

Answer

Use a SequentialPlanner with a stepwise strategy and explicit function parameter constraints

Answer

Use a ManualInvoke kernel with a sequential planner

Answer

Use a ParallelPlanner with a function-calling model

Question 7

You are implementing an agentic solution using Azure AI Agent Service. The agent needs to maintain conversation context across multiple turns. You configure the agent with a custom prompt that includes a 'system message' and 'few-shot examples'. However, after a few turns, the agent starts repeating the same responses. What is the most likely cause?

Accepted Answer

The 'max_context_length' is set too low, causing earlier turns to be truncated. When the 'max_context_length' is set too low, the agent's conversation history is truncated after a few turns, removing earlier user inputs and assistant responses. This loss of context causes the model to lose track of the conversation flow and revert to repeating responses, as it no longer has the full context to generate varied replies.

Answer

The 'max_tokens' parameter is set too low for the response

Answer

The agent is using a 'context recycling' feature that resets after each turn

Answer

The temperature is set too high, causing the model to become deterministic

Question 8

Your organization is deploying an agentic solution using Microsoft Copilot Studio. The agent must be able to escalate to a human agent when it cannot resolve a user's request. You need to ensure that the escalation includes the full conversation history. What should you configure?

Accepted Answer

Add a 'Transfer conversation' action and set it to include the full transcript. The 'Transfer conversation' action in Microsoft Copilot Studio is specifically designed to hand off the conversation to a human agent, and it includes an option to pass the full conversation transcript. This ensures the human agent has complete context, which is required for a seamless escalation. The other options do not provide this capability.

Answer

Add an 'End conversation' action and configure a fallback

Answer

Add a 'Create a ticket' action in the topic

Answer

Add a 'Start a new topic' action with context variables

Question 9

You are designing an agentic solution using Azure AI Agent Service with a custom skill that calls an external REST API. The API has rate limits: 100 requests per minute per client. You need to ensure the agent respects this limit without degrading user experience. Which approach should you take?

Accepted Answer

Use a token bucket rate limiter with a shared counter stored in Azure Cache for Redis. A token bucket rate limiter with a shared counter stored in Azure Cache for Redis provides a distributed, atomic mechanism to enforce a strict 100 requests per minute per client limit across multiple instances of the agent. This approach ensures that the rate limit is respected globally without introducing unnecessary delays when the limit is not reached, thus maintaining a responsive user experience.

Answer

Set a fixed delay of 600ms between each API call

Answer

Configure the skill to handle HTTP 429 responses with retry-after logic

Answer

Implement exponential backoff in the skill code

Question 10

You are building an agentic solution using Microsoft Semantic Kernel. The agent needs to orchestrate multiple plugins. One plugin returns a large dataset that exceeds the model's context window. What is the best way to handle this?

Accepted Answer

Use a summarization plugin to condense the data before passing it to the model. Semantic Kernel agents interact with LLMs that have fixed context windows. When a plugin returns data that exceeds this limit, the best practice is to use a summarization plugin (e.g., a built-in Semantic Kernel text summarizer or a custom one) to condense the data into a concise representation that fits within the model's token budget. This preserves the essential information without losing context or requiring manual truncation, which could discard critical details.

Answer

Split the data into multiple smaller API calls and combine results

Answer

Truncate the data to fit the context window

Answer

Configure the plugin to return only a subset of the data and mark the rest as sensitive