Microsoft AI-103 - Developing AI Apps and Agents on Azure

Agent1 requires memory because the business requirement explicitly states that the agent must retain conversation context and recall relevant information during future interactions. Microsoft Foundry Agent Service memory provides persistent knowledge that can be retained and retrieved across separate conversations, sessions, devices, and workflows. It can store meaningful information extracted from prior interactions, including user preferences, relevant facts, and summarized context, and then retrieve that information when personalizing later responses.

Short-term memory maintains context within the current conversation. Long-term memory supports continuity across future sessions, which directly addresses Contoso’s requirement. Memory scopes should be associated with individual user identities so that one customer’s stored information cannot be retrieved during another customer’s interaction.

Guardrails enforce safety or behavioral restrictions but do not provide conversation recall. Tools enable the agent to access external systems, while instructions define the agent’s role and operating boundaries. Neither tools nor instructions persist user-specific information across sessions.

The AI-103 Study Guide maps this requirement to choose appropriate memory, tool, and knowledge integration services and build agents that integrate retrieval, function calling, and conversation memory .

====================

Call client.videos.create() to submit the prompt and start a video-generation job against the model deployment. Video generation is asynchronous: the initial response represents a job and contains properties such as its unique ID and current status rather than immediately returning the completed video. Microsoft’s official Python example creates the job by using client.videos.create(model=..., prompt=...).

The application must then poll the job. After waiting between requests, client.videos.retrieve(video.id) obtains the latest state of that specific job. Polling continues until the status becomes completed, failed, or cancelled. list would enumerate jobs rather than refresh the selected job, while download_content is used only after successful completion to retrieve the generated video file.

To comply with Contoso’s prohibition on API keys, the client should use a Microsoft Entra bearer-token provider. The OpenAI Python client accepts this provider through its api_key constructor parameter even though no static API key is used.

Study Guide alignment: design and implement video-generation solutions and integrate generative workflows through Foundry SDKs.

A Retrieval Augmented Generation evaluator is required because Agent1 will retrieve product-sheet content and use it as grounding context when generating answers. Foundry RAG evaluators assess both retrieval quality and the quality of the final grounded response. This directly supports Contoso’s requirement that answers based on the product sheets be relevant, complete, and accurate.

The appropriate evaluation suite can include groundedness , which determines whether the answer is supported by the retrieved product information without fabrication; relevance , which evaluates whether the answer accurately and completely addresses the customer’s question; and response completeness , which measures whether critical expected information is missing. Retrieval evaluators can additionally determine whether the indexing and search pipeline returned context chunks relevant to the query.

A coherence evaluator assesses readability and logical organization but does not establish that claims are supported by storage1. A custom guardrail controls or blocks behavior rather than measuring RAG quality. Fine-tuning changes model behavior and does not evaluate whether retrieved documentation was selected and used correctly.

The AI-103 Study Guide includes evaluating models and applications for fabrications, relevance, quality, and safety , as well as monitoring grounding quality and retrieval relevance.

The correct answer is B. Personally Identifiable Information (PII) Detection . The case study states that Agent1 must never reveal customer information , even if a document containing customer data is added accidentally to the product sheet repository in storage1. This is a privacy and compliance control requirement, so the appropriate capability is PII Detection.

Azure Language PII Detection is a Foundry Tools capability that identifies, classifies, and redacts sensitive information across text, conversations, and native documents. Microsoft states that PII Detection can be used to implement privacy controls, reduce sensitive data exposure, and support compliance requirements. In this scenario, PII Detection should be applied to retrieved product-sheet content and generated responses so customer names, contact details, identifiers, and other sensitive values are not exposed to users.

Prompt Shields are important for a separate requirement: protecting Agent1 from malicious instructions hidden in documents or embedded text. Microsoft describes Prompt Shields for documents as protection against hidden instructions embedded in external content. However, the option that directly satisfies the requirement to prevent disclosure of customer information is PII Detection. Self-harm and violence filters address harmful-content categories, not privacy leakage.

The SSML uses three distinct neural voices: an en-US Jenny voice, an en-GB Ryan voice, and an en-US Christopher voice. Therefore, the output contains sentences produced by three different voices , making the first statement Yes .

The second statement is No . There are three voices but not three unique locale/accent configurations. Two voices use en-US while one uses en-GB. Consequently, the audio does not represent three different accents/locales.

The third statement is also No . The Christopher voice is wrapped inside an mstts:express-as element with an expressive style such as advertisement_upbeat, with a configured style degree. That instruction intentionally alters speaking style and therefore is not a neutral presentation.

SSML permits a single synthesis request to contain multiple < voice > elements, enabling multiple voices to appear in one output audio stream. SSML also supports style controls, pronunciation controls, pauses, pitch, speed, and other speech characteristics.

The AI-103 Study Guide includes text-to-speech for agentic interactions and speech modality configuration within text-analysis solutions. Microsoft Learn

====================

The solution does not meet the goal. Increasing max_tokens only raises the maximum number of tokens the model is allowed to generate. Microsoft’s Azure OpenAI reference defines max_tokens as the maximum number of tokens allowed for the generated answer, and the quota guidance notes that increasing it can help when responses are being truncated.

In this scenario, the problem is not described as output truncation. The required regulatory clauses are already present in the retrieved policy documents, but the agent omits them during summarization. That is a response completeness issue: Microsoft Foundry RAG evaluator guidance defines response completeness as the recall aspect of the response, meaning the response should not miss critical information compared with expected content or ground truth.

A larger token budget might permit a longer answer, but it does not force the model to identify, verify, or include each mandatory clause. It can also increase cost and latency. The appropriate control is a reflection or completeness verification pass that checks the draft against the retrieved policy clauses and regenerates or revises the response when required content is missing. Reference topics: RAG response completeness, model output limits, max_tokens, reflection, and response validation.

The correct capability is token usage . In Microsoft Foundry observability, token consumption is the primary signal for diagnosing model-cost changes when request volume is unchanged. Token usage lets you distinguish whether costs increased because prompts became larger, retrieved or tool-provided context expanded, responses became longer, or agent execution added more model calls. Microsoft Foundry monitoring dashboards track operational metrics such as token consumption, latency, error rates, and quality scores, and the agent monitoring dashboard is specifically intended to help analyze token usage, latency, success rates, and evaluation outcomes for production traffic.

This directly matches the scenario because the issue is not more traffic, but changed request or response characteristics. Input tokens reveal whether the prompt, chat history, grounding data, or tool outputs being sent to the model increased. Output tokens reveal whether the model is generating longer completions. Expanded tool usage can also increase cost indirectly by adding more tool results, intermediate calls, and context into subsequent model requests; Foundry tracing and observability capture tool usage and token consumption for agent runs.

Evaluation metrics assess response quality and safety, not cost drivers. Latency identifies performance delays, and run success rate measures reliability. Reference topics: Microsoft Foundry observability, agent monitoring dashboard, token consumption, cost analysis, tool usage, and production monitoring.

Comprehensive and Detailed 150 to 250 words of Explanation From Azure AI Apps and Agents Developer/Course Guide/topics:

Use a conversation . A conversation is the durable runtime container that maintains the sequence of interactions associated with a particular case or user flow.

The application can persist the conversation ID together with the support-case ID. When the same case resumes later—even in another application session—the application reuses that conversation ID. The runtime can then continue with the prior interaction history rather than requiring the application to reconstruct it manually.

The stored history can include user messages, assistant responses, tool-call requests, and tool outputs, which is precisely what the question requires.

An agent defines reusable behavior such as the model, instructions, and tools. Many independent conversations can use the same agent, so the agent ID is not the state container for an individual support case.

A response represents an individual model/run result, while an output item is one item produced during response processing. Neither represents the complete durable interaction history.

This separation between durable conversational state and reusable agent configuration is central to Foundry ' s agent runtime model.

Study Guide alignment: define conversation-tracking approaches and build agents that integrate retrieval, function calling, and conversation memory . Microsoft Learn

====================

Comprehensive and Detailed 150 to 250 words of Explanation From Azure AI Apps and Agents Developer/Course Guide/topics:

Use the Azure Content Understanding skill. Content Understanding combines multimodal document processing, layout understanding, semantic extraction, image/figure handling, and structured document interpretation in a single content-processing capability.

For citation-rich RAG scenarios, the generated output can retain page and spatial information associated with source content, allowing downstream applications to connect returned text and visual content to page-level locations and bounding regions. Its richer document analysis is also designed to handle complex structures such as tables, figures, and content relationships that span document pages.

The ordinary Document Extraction skill primarily performs document cracking and basic extraction. Document Layout focuses on layout structure but does not provide the same complete multimodal semantic processing required by the question. GenAI Prompt applies generative processing to already available content rather than serving as the primary multimodal document-extraction skill.

Content Understanding therefore minimizes the need to compose multiple separate skills for OCR, image extraction, layout, and semantic document processing.

The AI-103 Study Guide explicitly covers multimodal pipelines that combine OCR, layout analysis, and field extraction and producing clean grounded representations for agents and RAG by using Content Understanding. Microsoft Learn

====================

Change the model domain to a compact domain.

Retrain the model.

Export the model.

Comprehensive and Detailed 150 to 250 words of Explanation From Azure AI Apps and Agents Developer/Course Guide/topics:

A Custom Vision model can be packaged for local mobile inference only when it is trained by using an exportable compact domain .

The current Retail-domain model is intended for its hosted Custom Vision scenario and cannot simply be copied into an Android application. Therefore, first change the model/project domain to the appropriate compact object-detection domain, such as General (compact) where applicable.

Second, retrain the model . The existing Retail iteration was trained using a different underlying model/domain configuration. Retraining creates a new compact iteration that is optimized for export and edge inference.

Third, export the compact model using an Android-compatible format supported by Custom Vision and the app ' s inference runtime. Depending on the available project/export options, this can involve formats such as TensorFlow/TensorFlow Lite or ONNX.

Testing is useful before production deployment but does not replace any of the three steps required to transform the hosted Retail model into an exportable mobile model.

The correct lifecycle is therefore:

change domain → retrain → export

Study Guide alignment: implement object/component detection and prepare vision solutions for appropriate local or mobile execution environments .