Main Flashcards
What is the primary purpose of Azure AI Content Safety?
Azure AI Content Safety is a comprehensive solution designed to detect harmful user-generated and AI-generated content in applications and services
Which Azure AI service would you select for detecting offensive content in text?
Azure AI Content Safety is the appropriate service for detecting offensive content in text with multiple severity levels.
What are the main capabilities of Azure AI Content Safety for text analysis?
Azure AI Content Safety can scan text for sexual content
How many languages does Azure AI Content Safety support?
Azure AI Content Safety supports more than 100 languages and is specifically trained on English
What is the difference between Azure Content Moderator and Azure AI Content Safety?
Azure Content Moderator is deprecated as of February 2024 and will be retired by February 2027
What is the weight percentage for “Plan and manage an Azure AI solution” in the AI-102 exam?
15-20% of the exam covers planning and managing an Azure AI solution.
What is the weight percentage for “Implement content moderation solutions” in the AI-102 exam?
10-15% of the exam covers implementing content moderation solutions.
What is the weight percentage for “Implement computer vision solutions” in the AI-102 exam?
15-20% of the exam covers implementing computer vision solutions.
What is the weight percentage for “Implement natural language processing solutions” in the AI-102 exam?
30-35% of the exam covers implementing natural language processing solutions.
What is the weight percentage for “Implement knowledge mining and document intelligence solutions” in the AI-102 exam?
10-15% of the exam covers implementing knowledge mining and document intelligence solutions.
What is the weight percentage for “Implement generative AI solutions” in the AI-102 exam?
10-15% of the exam covers implementing generative AI solutions.
What are the two programming languages an Azure AI engineer should have experience with?
Python and C#.
What type of APIs should an Azure AI engineer be able to use to build secure AI solutions?
Representational State Transfer (REST) APIs and SDKs.
What is the primary responsibility of an Azure AI engineer?
Building
What phases of AI solutions development does an Azure AI engineer participate in?
Requirements definition and design
Which Azure AI service would you use for image classification?
Azure AI Vision service is used for image classification.
Which Azure AI service would you use for object detection in images?
Azure AI Vision service is used for object detection in images.
Which Azure AI service would you use for extracting text from images?
Azure AI Vision service with OCR capability is used for extracting text from images.
Which Azure AI service would you use for detecting faces in images?
Azure AI Vision service is used for detecting faces in images.
Which Azure AI service would you use for analyzing sentiment in text?
Azure AI Language service is used for analyzing sentiment in text.
Which Azure AI service would you use for extracting key phrases from text?
Azure AI Language service is used for extracting key phrases from text.
Which Azure AI service would you use for entity recognition in text?
Azure AI Language service is used for entity recognition in text.
Which Azure AI service would you use for language detection?
Azure AI Language service is used for language detection.
Which Azure AI service would you use for speech-to-text conversion?
Azure AI Speech service is used for speech-to-text conversion.
Which Azure AI service would you use for text-to-speech conversion?
Azure AI Speech service is used for text-to-speech conversion.
Which Azure AI service would you use for translating text between languages?
Azure AI Translator service is used for translating text between languages.
Which Azure AI service would you use for knowledge mining?
Azure AI Search service is used for knowledge mining.
Which Azure AI service would you use for document intelligence?
Azure AI Document Intelligence service is used for document intelligence.
Which Azure AI service would you use for generating natural language content?
Azure OpenAI Service is used for generating natural language content.
Which Azure AI service would you use for generating images from text descriptions?
Azure OpenAI Service with DALL-E model is used for generating images from text descriptions.
What is the purpose of Azure AI Video Indexer?
Azure AI Video Indexer is used to extract insights from videos or live streams.
What is Azure AI Vision Spatial Analysis used for?
Azure AI Vision Spatial Analysis is used to detect the presence and movement of people in video.
What is SSML in the context of Azure AI Speech?
Speech Synthesis Markup Language (SSML) is used to improve text-to-speech by controlling aspects like pronunciation
What is a custom question answering solution in Azure AI Language?
A custom question answering solution allows you to create a knowledge base of questions and answers that can be queried using natural language.
What is an Azure AI Search skillset?
A skillset in Azure AI Search is a collection of cognitive skills that extract and enrich data during indexing.
What is a Knowledge Store projection in Azure AI Search?
A Knowledge Store projection is a way to save enriched documents in Azure Storage as files
What is a composed document intelligence model?
A composed document intelligence model combines multiple document intelligence models to extract data from complex documents.
What is prompt engineering in the context of Azure OpenAI?
Prompt engineering involves crafting effective prompts to improve the responses generated by Azure OpenAI models.
What is fine-tuning in the context of Azure OpenAI?
Fine-tuning is the process of further training an Azure OpenAI model on specific data to improve its performance for particular tasks.
How do you create an Azure AI resource?
You create an Azure AI resource through the Azure portal
How do you determine a default endpoint for an Azure AI service?
The default endpoint for an Azure AI service is typically in the format https://{resource-name}.{region}.api.cognitive.microsoft.com and can be found in the Azure portal under the resource’s Keys and Endpoint section.
How do you integrate Azure AI services into a CI/CD pipeline?
You integrate Azure AI services into a CI/CD pipeline by using Azure DevOps or GitHub Actions to automate the deployment and testing of AI models and services.
How do you plan and implement a container deployment for Azure AI services?
You plan and implement a container deployment by selecting container-supported services
How do you configure diagnostic logging for Azure AI services?
You configure diagnostic logging by enabling Azure Monitor diagnostics for the AI resource and specifying log categories and destinations like Log Analytics
How do you monitor an Azure AI resource?
You monitor an Azure AI resource using Azure Monitor to track metrics like request count
How do you manage costs for Azure AI services?
You manage costs by selecting appropriate pricing tiers
How do you manage account keys for Azure AI services?
You manage account keys by regularly regenerating them
How do you protect account keys using Azure Key Vault?
You protect account keys by storing them in Azure Key Vault and accessing them securely in applications using Key Vault references or managed identities.
How do you manage authentication for an Azure AI service resource?
You manage authentication by using API keys for simple scenarios or implementing OAuth 2.0 with Azure AD for more secure enterprise applications.
How do you manage private communications for Azure AI services?
You manage private communications by implementing Private Link or Virtual Network service endpoints to restrict network access to your AI resources.
How do you implement a text moderation solution with Azure AI Content Safety?
You implement a text moderation solution by creating an Azure AI Content Safety resource
How do you implement an image moderation solution with Azure AI Content Safety?
You implement an image moderation solution by creating an Azure AI Content Safety resource
How do you select visual features to meet image processing requirements?
You select visual features by identifying the specific information needed from images (objects
How do you detect objects in images and generate image tags?
You detect objects and generate tags by using the Azure AI Vision service’s object detection and tagging capabilities through its API or SDK.
How do you include image analysis features in an image processing request?
You include image analysis features by specifying the desired visual features (like Categories
How do you interpret image processing responses from Azure AI Vision?
You interpret responses by parsing the JSON output to extract relevant information like detected objects
How do you extract text from images using Azure AI Vision?
You extract text using the OCR (Optical Character Recognition) capability of Azure AI Vision by calling the Read API and processing the returned text content.
How do you convert handwritten text using Azure AI Vision?
You convert handwritten text using the OCR capability of Azure AI Vision
How do you choose between image classification and object detection models?
You choose image classification when you need to categorize entire images
How do you label images for custom vision models?
You label images by uploading them to the Custom Vision portal or using the SDK
How do you train a custom image classification model?
You train a custom image classification model by uploading labeled images to Custom Vision
How do you train a custom object detection model?
You train a custom object detection model by uploading images with bounding box annotations to Custom Vision
How do you evaluate custom vision model metrics?
You evaluate metrics by reviewing precision
How do you publish a custom vision model?
You publish a model by selecting the trained iteration in Custom Vision and publishing it to an endpoint with a prediction resource.
How do you consume a custom vision model?
You consume a model by making HTTP requests to the published prediction endpoint using the provided API key
How do you use Azure AI Video Indexer to extract insights from a video?
You use Video Indexer by uploading videos to the service
How do you use Azure AI Vision Spatial Analysis to detect people in video?
You use Spatial Analysis by deploying specialized containers that process video streams to detect people and their movements in physical spaces.
How do you extract key phrases from text using Azure AI Language?
You extract key phrases by sending text to the Key Phrase Extraction API of Azure AI Language
How do you extract entities from text using Azure AI Language?
You extract entities by sending text to the Named Entity Recognition API of Azure AI Language
How do you determine sentiment of text using Azure AI Language?
You determine sentiment by sending text to the Sentiment Analysis API of Azure AI Language
How do you detect the language used in text?
You detect language by sending text to the Language Detection API of Azure AI Language
How do you detect personally identifiable information (PII) in text?
You detect PII by using the PII Detection capability of Azure AI Language
How do you implement text-to-speech using Azure AI Speech?
You implement text-to-speech by sending text to the Speech service’s text-to-speech API
How do you implement speech-to-text using Azure AI Speech?
You implement speech-to-text by sending audio to the Speech service’s speech-to-text API
How do you improve text-to-speech using SSML?
You improve text-to-speech by formatting your input text with SSML tags that control aspects like pronunciation
How do you implement custom speech solutions?
You implement custom speech solutions by creating acoustic and language models trained on your specific data to improve recognition accuracy for your domain.
How do you implement intent recognition with Azure AI Speech?
You implement intent recognition by integrating Speech service with Language Understanding to recognize both speech and the user’s intent.
How do you implement keyword recognition with Azure AI Speech?
You implement keyword recognition by creating custom keyword models that can detect specific trigger words or phrases in audio streams.
How do you translate text using the Azure AI Translator service?
You translate text by sending content to the Translator API with source and target language parameters to receive the translated output.
How do you implement custom translation?
You implement custom translation by training models with your parallel text data in the Custom Translator portal and deploying them for use.
How do you translate speech-to-speech using Azure AI Speech?
You translate speech-to-speech by using the Speech Translation API
How do you translate speech-to-text using Azure AI Speech?
You translate speech-to-text by using the Speech Translation API
How do you translate to multiple languages simultaneously?
You translate to multiple languages by specifying multiple target languages in a single Translator API request to receive translations for all requested languages.
How do you create intents in a language understanding model?
You create intents by defining categories of user queries in the Language Understanding service and providing example utterances for each intent.
How do you add utterances to intents in a language understanding model?
You add utterances by providing example phrases that users might say for each intent to help the model recognize similar expressions.
How do you create entities in a language understanding model?
You create entities by defining data you want to extract from utterances
How do you train a language understanding model?
You train a model by submitting your defined intents
How do you evaluate a language understanding model?
You evaluate a model by reviewing performance metrics like intent recognition accuracy and entity extraction precision.
How do you deploy a language understanding model?
You deploy a model by publishing it to a production endpoint that applications can query for predictions.
How do you test a language understanding model?
You test a model by submitting sample utterances to the deployed endpoint and verifying that intents and entities are correctly recognized.
How do you optimize a language understanding model?
You optimize a model by reviewing prediction results
How do you consume a language model from a client application?
You consume a model by making HTTP requests to the published endpoint using the provided API key
How do you backup language understanding models?
You backup models by exporting them as JSON files that contain all intents
How do you recover language understanding models?
You recover models by importing previously exported JSON files to recreate the language understanding model.
How do you create a custom question answering project?
You create a project in the Language Studio by defining its name
How do you add question-and-answer pairs manually?
You add Q&A pairs by entering questions and their corresponding answers in the Language Studio interface.
How do you import sources to a question answering solution?
You import sources by adding URLs
How do you train and test a knowledge base?
You train a knowledge base by saving and building it
How do you publish a knowledge base?
You publish a knowledge base by deploying it to an endpoint that applications can query for answers.
How do you create a multi-turn conversation in question answering?
You create multi-turn conversations by defining follow-up prompts for specific questions to guide users through a conversational flow.
How do you add alternate phrasing to a knowledge base?
You add alternate phrasing by providing multiple ways to ask the same question
How do you add chit-chat to a knowledge base?
You add chit-chat by importing predefined personality datasets that provide responses to common conversational queries.
How do you export a knowledge base?
You export a knowledge base by downloading its content as a TSV or JSON file for backup or transfer.
How do you create a multi-language question answering solution?
You create a multi-language solution by creating separate knowledge bases for each language and connecting them through a routing mechanism.
How do you provision an Azure AI Search resource?
You provision a resource by creating it in the Azure portal
How do you create data sources in Azure AI Search?
You create data sources by defining connections to supported data repositories like Azure SQL Database
How do you create an index in Azure AI Search?
You create an index by defining its schema with fields
How do you define a skillset in Azure AI Search?
You define a skillset by specifying a collection of cognitive skills that will process and enrich your data during indexing.
How do you implement custom skills in Azure AI Search?
You implement custom skills by creating web APIs that follow the skillset interface and including them in your skillset definition.
How do you create and run an indexer in Azure AI Search?
You create an indexer by defining a process that connects a data source to an index
How do you query an index in Azure AI Search?
You query an index by sending search requests to the search API with parameters for text search
How do you use syntax in Azure AI Search queries?
You use syntax like Lucene query syntax or simple query syntax to create complex search expressions with operators and wildcards.
How do you implement sorting in Azure AI Search queries?
You implement sorting by specifying one or more fields to sort by and the sort direction (ascending or descending).
How do you implement filtering in Azure AI Search queries?
You implement filtering by adding OData $filter expressions to limit results based on field values.
How do you use wildcards in Azure AI Search queries?
You use wildcards like * and ? in search terms to match patterns at the beginning
How do you manage Knowledge Store projections?
You manage projections by defining how enriched data should be saved to Azure Storage as files
How do you provision a Document Intelligence resource?
You provision a resource by creating it in the Azure portal
How do you use prebuilt models to extract data from documents?
You use prebuilt models by sending documents to the appropriate model endpoint (like receipt
How do you implement a custom document intelligence model?
You implement a custom model by collecting sample documents
How do you train a custom document intelligence model?
You train a custom model by uploading labeled documents to the Document Intelligence Studio and initiating the training process.
How do you test a custom document intelligence model?
You test a model by analyzing new documents and verifying that fields are correctly extracted with acceptable confidence scores.
How do you publish a custom document intelligence model?
You publish a model by deploying it to an endpoint that applications can use to analyze documents.
How do you create a composed document intelligence model?
You create a composed model by combining multiple models (prebuilt or custom) to extract data from complex documents with varied layouts.
How do you implement a document intelligence model as a custom Azure AI Search skill?
You implement it as a custom skill by creating a web API that uses the Document Intelligence SDK and integrating it into an Azure AI Search skillset.
How do you provision an Azure OpenAI Service resource?
You provision a resource by creating it in the Azure portal
How do you select and deploy an Azure OpenAI model?
You select a model by choosing from available options (like GPT-4
How do you submit prompts to generate natural language?
You submit prompts by sending requests to the Completions or Chat Completions API with your text prompt and model parameters.
How do you submit prompts to generate code?
You submit prompts by sending requests to the Completions or Chat Completions API with code-related instructions and appropriate model parameters.
How do you use the DALL-E model to generate images?
You use DALL-E by sending requests to the Images API with text descriptions of the images you want to generate.
How do you use Azure OpenAI APIs to submit prompts and receive responses?
You use the APIs by making HTTP requests to your deployed model’s endpoint with your API key and prompt data.
How do you use large multimodal models in Azure OpenAI?
You use multimodal models by sending requests that can include both text and image inputs to generate responses based on multiple types of content.
How do you configure parameters to control generative behavior?
You configure parameters like temperature
How do you apply prompt engineering techniques?
You apply prompt engineering by crafting effective prompts with clear instructions
How do you use your own data with an Azure OpenAI model?
You use your own data by implementing Azure AI Search integration or other retrieval methods to provide relevant context from your documents.
How do you fine-tune an Azure OpenAI model?
You fine-tune a model by preparing a dataset of examples and using the fine-tuning API to adapt the model to your specific use case.
What is the purpose of Responsible AI principles in Azure AI solutions?
Responsible AI principles ensure AI systems are fair
How do you plan for a solution that meets Responsible AI principles?
You plan by conducting impact assessments
What is the difference between image classification and object detection?
Image classification assigns labels to entire images
What is OCR in the context of Azure AI Vision?
OCR (Optical Character Recognition) is the process of extracting text from images
What is the purpose of custom vision models?
Custom vision models allow you to train specialized image classification or object detection models on your specific visual content.
What metrics are used to evaluate custom vision models?
Precision (accuracy of positive predictions)
What is the difference between key phrase extraction and entity recognition?
Key phrase extraction identifies important phrases in text
What is sentiment analysis in Azure AI Language?
Sentiment analysis determines the emotional tone of text
What is the purpose of language detection in Azure AI Language?
Language detection identifies which language is used in a text document
What is PII detection in Azure AI Language?
PII detection identifies personally identifiable information in text
What is the difference between text-to-speech and speech-to-text?
Text-to-speech converts written text into spoken audio
What is SSML and why is it used?
Speech Synthesis Markup Language (SSML) is an XML-based markup language used to control aspects of speech synthesis like pronunciation
What is intent recognition in the context of speech processing?
Intent recognition identifies the purpose or goal behind a user’s spoken request
What is keyword recognition in Azure AI Speech?
Keyword recognition detects specific trigger words or phrases in audio streams
What is the difference between translating text and translating speech?
Text translation converts written text between languages
What is custom translation in Azure AI Translator?
Custom translation allows you to train translation models on your domain-specific parallel text data to improve translation quality for your content.
What are intents in a language understanding model?
Intents are categories that represent the purpose or goal behind user queries or commands.
What are utterances in a language understanding model?
Utterances are example phrases that users might say to express a particular intent.
What are entities in a language understanding model?
Entities are pieces of data you want to extract from utterances
What is the difference between training and evaluating a language understanding model?
Training builds the model based on provided intents
What is a knowledge base in question answering?
A knowledge base is a collection of question-and-answer pairs that can be queried using natural language to provide relevant answers.
What is multi-turn conversation in question answering?
Multi-turn conversation allows for follow-up questions and context-aware responses in a dialog flow.
What is chit-chat in a knowledge base?
Chit-chat provides responses to common conversational queries that aren’t directly related to the primary purpose of the knowledge base.
What is the purpose of Azure AI Search?
Azure AI Search provides full-text search capabilities with AI-powered content understanding for applications.
What is a data source in Azure AI Search?
A data source is a connection to a repository like Azure SQL Database
What is an index in Azure AI Search?
An index is a searchable collection of documents with a defined schema that determines how the data is stored and searched.
What is a skillset in Azure AI Search?
A skillset is a collection of cognitive skills that extract and enrich data during the indexing process.
What is an indexer in Azure AI Search?
An indexer is a process that connects a data source to an index
What is the difference between syntax
sorting
What is a Knowledge Store projection?
A Knowledge Store projection saves enriched data from the indexing process to Azure Storage in various formats.
What is Document Intelligence in Azure AI?
Document Intelligence extracts structured data from documents like forms
What is the difference between prebuilt and custom document intelligence models?
Prebuilt models are ready-to-use for common document types
What is a composed document intelligence model?
A composed model combines multiple document intelligence models to extract data from complex documents with varied layouts.
What is Azure OpenAI Service?
Azure OpenAI Service provides access to large language models like GPT-4 and DALL-E with Azure security and compliance features.
What is the difference between generating natural language and generating code?
Natural language generation creates human-like text for content
What is DALL-E in Azure OpenAI?
DALL-E is a model that generates images from text descriptions.
What are multimodal models in Azure OpenAI?
Multimodal models can process and generate content based on multiple types of inputs
What parameters control generative behavior in Azure OpenAI?
Parameters like temperature
What is prompt engineering?
Prompt engineering is the practice of crafting effective prompts to guide AI models toward desired outputs.
What is fine-tuning in Azure OpenAI?
Fine-tuning adapts a pre-trained model to specific use cases by training it on additional examples.
How do you implement a text moderation workflow with Azure AI Content Safety?
You implement a workflow by integrating the Content Safety API into your content processing pipeline
How do you handle multiple languages in content moderation?
You handle multiple languages by using Azure AI Content Safety’s multilingual capabilities and specifying the language parameter when available.
How do you implement real-time image moderation?
You implement real-time moderation by integrating the Content Safety API into your upload or display process with appropriate error handling and fallback mechanisms.
How do you balance false positives and false negatives in content moderation?
You balance detection errors by adjusting severity thresholds based on your application’s requirements and monitoring moderation results.
How do you implement batch processing for image analysis?
You implement batch processing by submitting multiple images in a single request or using asynchronous processing for large volumes.
How do you handle low-quality or blurry images in computer vision?
You handle low-quality images by implementing pre-processing techniques
How do you implement a solution that combines image classification and object detection?
You implement a combined solution by using both capabilities in sequence or parallel and integrating their results based on your application needs.
How do you optimize computer vision models for edge devices?
You optimize models by using quantization
How do you implement a multilingual text analysis solution?
You implement a multilingual solution by detecting the language first and then routing to appropriate language-specific processing or using multilingual models.
How do you handle domain-specific terminology in text analysis?
You handle domain-specific terminology by using custom models
How do you implement a solution that combines multiple text analysis capabilities?
You implement a combined solution by calling multiple APIs in sequence or parallel and integrating their results based on your application needs.
How do you handle speech recognition in noisy environments?
You handle noisy environments by using acoustic models trained on diverse audio conditions
How do you implement a multilingual speech solution?
You implement a multilingual solution by detecting the language or allowing users to select it
How do you optimize speech recognition for specific domains?
You optimize domain recognition by training custom speech models with domain-specific vocabulary and acoustic data.
How do you implement a solution that combines speech and text analysis?
You implement a combined solution by converting speech to text first
How do you implement a cross-language communication solution?
You implement cross-language communication by combining speech recognition
How do you optimize a question answering solution for domain-specific queries?
You optimize for domain-specific queries by adding comprehensive Q&A pairs
How do you implement a solution that combines question answering with other language capabilities?
You implement a combined solution by using question answering for knowledge retrieval and other language capabilities for additional processing.
How do you implement incremental indexing in Azure AI Search?
You implement incremental indexing by using change detection mechanisms like high watermarks or change tracking in your data source.
How do you optimize search relevance in Azure AI Search?
You optimize relevance by configuring scoring profiles
How do you implement faceted navigation in search results?
You implement faceted navigation by defining facetable fields in your index and requesting facet counts in your queries.
How do you implement semantic search in Azure AI Search?
You implement semantic search by enabling the semantic configuration on your index and using semantic ranking in your queries.
How do you handle document security in Azure AI Search?
You handle security by implementing security trimming with user identity information and access control filters.
How do you optimize document processing for various document types?
You optimize processing by selecting appropriate prebuilt models
How do you implement a solution that extracts data from handwritten forms?
You implement handwritten extraction by using Document Intelligence with models that support handwriting recognition.
How do you handle document processing errors and low confidence extractions?
You handle errors by implementing confidence thresholds
How do you implement a solution that combines document intelligence with other AI capabilities?
You implement a combined solution by using document intelligence for extraction and other AI services for further analysis of the extracted data.
How do you implement responsible AI practices in generative AI solutions?
You implement responsible AI by using content filtering
How do you optimize token usage in Azure OpenAI?
You optimize token usage by crafting efficient prompts
How do you implement a solution that combines generative AI with other Azure AI services?
You implement a combined solution by using generative AI for content creation and other AI services for analysis
How do you implement a retrieval-augmented generation (RAG) pattern?
You implement RAG by combining Azure AI Search for knowledge retrieval with Azure OpenAI for generating contextually relevant responses.
How do you handle sensitive information in generative AI inputs and outputs?
You handle sensitive information by implementing PII detection
What is the Azure AI Vision service used for?
Azure AI Vision is used for image analysis tasks like object detection
What is the Azure AI Language service used for?
Azure AI Language is used for text analysis tasks like sentiment analysis
What is the Azure AI Speech service used for?
Azure AI Speech is used for speech processing tasks like speech-to-text
What is the Azure AI Translator service used for?
Azure AI Translator is used for translating text and documents between languages.
What is Azure AI Search used for?
Azure AI Search is used for implementing search capabilities with AI-powered content understanding and knowledge mining.
What is Azure AI Document Intelligence used for?
Azure AI Document Intelligence is used for extracting structured data from documents like forms
What is Azure OpenAI Service used for?
Azure OpenAI Service is used for generating natural language
What is the difference between Azure AI Vision and Custom Vision?
Azure AI Vision provides general image analysis capabilities
What is the difference between Azure AI Language and Language Understanding?
Azure AI Language is the broader service that includes Language Understanding as one of its capabilities for intent recognition and entity extraction.
What is the difference between Azure AI Search and traditional search engines?
Azure AI Search includes AI-powered content understanding
What is the difference between Azure AI Document Intelligence and OCR?
Azure AI Document Intelligence goes beyond OCR by not only extracting text but also understanding document structure and extracting specific fields with their relationships.
What is the difference between Azure OpenAI Service and OpenAI’s direct offerings?
Azure OpenAI Service provides OpenAI models with Azure’s security
