AI-102 Exam Dumps - Designing and Implementing a Microsoft Azure AI Solution

To quickly improve a GPT-3.5 chatbot’s response quality with minimal development effort:

Provide grounding content (B): Ground the model with authoritative data (e.g., the travel agent’s policies, inventory, fares) using Azure OpenAI on your data or a lightweight RAG pattern. Grounding reduces hallucinations and makes answers more accurate and context-aware without retraining the model. This typically involves configuring a data connection (Azure Cognitive Search/Blob) and augmenting prompts—far less effort than training.

Add sample request/response pairs (C): Use few-shot examples directly in the system/user messages to demonstrate ideal answers, style, and formatting. Few-shot prompting is a core prompt-engineering technique that meaningfully boosts quality with minimal code changes or infrastructure.

Options A, D, and E involve model training or retraining (fine-tuning or building a custom LLM), which require data preparation, experimentation, monitoring, and higher cost/complexity—contrary to the “minimize development effort” constraint.

Microsoft References

Prompt engineering with examples (few-shot) to improve output quality:https://learn.microsoft.com/azure/ai-services/openai/concepts/prompt-engineering

Azure OpenAI “on your data” (grounding to improve relevance and accuracy):https://learn.microsoft.com/azure/ai-services/openai/concepts/use-your-data

Guidance comparing prompt engineering/grounding with fine-tuning and when to use each:https://learn.microsoft.com/azure/ai-services/openai/concepts/fine-tuning

How-to: Fine-tuning (shows added complexity vs. prompt techniques):https://learn.microsoft.com/azure/ai-services/openai/how-to/fine-tuning

To create a single multi-service Cognitive Services resource (now Azure AI Services) that provides one key/endpoint for multiple services, consolidates billing, and allows adding services like Text Analytics (sentiment analysis) and Computer Vision (OCR) from the same resource, you must:

Use the Azure Resource Manager (ARM) REST API with PUT to create or update the account:PUT /subscriptions/{subId}/resourceGroups/{rg}/providers/Microsoft.CognitiveServices/accounts/{name}?api-version=2017-04-18

Set "kind": "CognitiveServices" to create the multi-service account (not a single-service kind like ComputerVision or TextAnalytics).

This satisfies the requirements: single key/endpoint, consolidated billing, and future Computer Vision use.

Microsoft Azure AI Solution References

ARM REST for Cognitive Services accounts (create/update via PUT, multi-service "kind": "CognitiveServices").https://learn.microsoft.com/azure/ai-services/managem ent/resource-manager-create-account

Azure AI Services multi-service resource (single endpoint/key across services).https://learn.microsoft.com/azure/ai-services/multi-service-resource

To make the chatbot respond only when the answer’s confidence is ≥ 95%, you must set the request body property confidenceScoreThreshold to 0.95 in the Query knowledge bases operation of Custom question answering. The REST API for querying a Custom question answering project is:

POST {Endpoint}/language/:query-knowledgebases?projectName={projectName}&deploymentName={deploymentName}&api-version=2021-10-01

Here, projectName identifies the Custom question answering project you’re querying (the screenshot’s placeholder ), while deploymentName specifies which deployment (for example, production) to use. The request body includes confidenceScoreThreshold (0–1) to filter out answers below that score; setting it to 0.95 enforces the requirement.

Example snippet:

curl -X POST "$LANGUAGE_ENDPOINT/language/:query-knowledgebases?projectName=&deploymentName=production&api-version=2021-10-01" \

-H "Ocp-Apim-Subscription-Key: $LANGUAGE_KEY" \

-H "Content-Type: application/json" \

-d '{

"question": "How much energy does my phone have left?",

"confidenceScoreThreshold": 0.95

}'

References

REST API: Question Answering – Get Answers (shows endpoint with projectName/deploymentName and the confidenceScoreThreshold request property). Microsoft Learn

REST API (newer version) parameter list for projectName/deploymentName. Microsoft Learn

Box 1: Cognitive Service User

Ensure that the members of a group named Management-Accountants can approve the FAQs.

Approve=publish.

Cognitive Service User (read/write/publish): API permissions: All access to Cognitive Services resource except for ability to:

1. Add new members to roles.

2. Create new resources.

Box 2: Cognitive Services QnA Maker Editor

Ensure that the members of a group named Consultant-Accountants can create and amend the FAQs.

QnA Maker Editor: API permissions:

1. Create KB API

2. Update KB API

3. Replace KB API

4. Replace Alterations

5. "Train API" [in new service model v5]

Box 3: Cognitive Services QnA Maker Read

Ensure that the members of a group named the Agent-CustomerServices can browse the FAQs.

QnA Maker Read: API Permissions:

1. Download KB API

2. List KBs for user API

3. Get Knowledge base details

4. Download Alterations

Generate Answer

Box 1: "categories": ["Locations", "Persons", "Organizations"],

Locations, Persons, Organizations are in the outputs.

Scenario: Contoso plans to develop a searchable knowledgebase of all the intellectual property

Note: The categories parameter is an array of categories that should be extracted. Possible category types: "Person", "Location", "Organization", "Quantity", "Datetime", "URL", "Email". If no category is provided, all types are returned.

Box 2: {"name": " entities"}

The include wikis, so should include entities in the outputs.

Note: entities is an array of complex types that contains rich information about the entities extracted from text, with the following fields

name (the actual entity name. This represents a "normalized" form)

wikipediaId

wikipediaLanguage

wikipediaUrl (a link to Wikipedia page for the entity)

etc.

The requirements state:

All AI solution responses must have a confidence score ≥ 70%.

If the response confidence score is < 70%, the response must be improved by human input.

Members of the Consultant-Bookkeeper group must be able to process financial documents, which includes performing manual review when the AI confidence is below threshold.

Looking at the provided JSON:

"fields": {

"ReceiptType": {

"type": "string",

"valueString": "Itemized",

"confidence": 0.672

},

"MerchantName": {

"type": "string",

"valueString": "Tailwind",

"confidence": 0.913

}

}

ReceiptType.confidence = 0.672 → this is below 0.7.

MerchantName.confidence = 0.913 → this is above 0.7.

Therefore, the correct condition for triggering manual review is:

documentResults.fields..confidence < 0.7

Now, evaluating the options:

A. documentResults.docType == "prebuilt:receipt"

Always true for receipts, does not check confidence. Not correct.

B. documentResults.fields.*.confidence < 0.7

This is the correct general expression: trigger manual review whenever any field confidence is below 0.7.

C. documentResults.fields.ReceiptType.confidence > 0.7

This would bypass manual review when ReceiptType has high confidence. The requirement is to trigger review when confidence < 0.7, so this is the opposite.

D. documentResults.fields.MerchantName.confidence < 0.7

This only checks one field (MerchantName). In the JSON, MerchantName has confidence 0.913 (>0.7), so this condition would not trigger, but ReceiptType clearly needs review. Too narrow.

Correct Answer: B. documentResults.fields.*.confidence < 0.7

Azure AI Document Intelligence – Confidence scores

Human-in-the-loop for Document Intelligence

Microsoft References

The wiki contains text in English, French and Portuguese.

Scenario: All planned projects must support English, French, and Portuguese.

The Document Extraction skill extracts content from a file within the enrichment pipeline. This allows you to take advantage of the document extraction step that normally happens before the skillset execution with files that may be generated by other skills.

Note: The Translator Text API will be used to determine the from language. The Language detection skill is not required.

Scenario: All planned projects must support English, French, and Portuguese.

If a bot uses multiple LUIS models and QnA Maker knowledge bases (knowledge bases), you can use the Dispatch tool to determine which LUIS model or QnA Maker knowledge base best matches the user input. The dispatch tool does this by creating a single LUIS app to route user input to the correct model.