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

Designing and Implementing a Microsoft Azure AI Solution (AI-102) exam is designed for individuals having skills in building, managing, and deploying AI solutions that support Azure Cognitive Services, Azure Cognitive Search, and Microsoft Bot Framework. In this exam, candidates are responsible for participating in all phases of AI solutions development whether for requirements definition or in designing to development, deployment, maintenance, performance tuning, and monitoring. Lastly, candidates in this will work with solution architects for translating their vision and with data scientists, data engineers, IoT specialists, and AI developers for building complete end-to-end AI solutions. Passing the AI-102 exam will help individuals in becoming Microsoft Certified Azure AI Engineer Associate.

Knowledge Requirement for the AI-102 Exam

• Firstly, the candidates must have proficiency in C#, Python, or JavaScript and have the ability to use REST-based APIs and SDKs for building computer vision, natural language processing, knowledge mining, and conversational AI solutions on Azure. 
• Secondly, they must have familiarity with the Azure AI portfolio components and the available data storage options. 
• Lastly, they should understand and have the ability to apply responsible AI principles.

Microsoft AI-102 Exam Learning Path

Microsoft provides access to its learning path for the AI-102 certification exams. This learning path for the AI-102 exam consists of topics covering modules that will help candidates to understand the concepts in a step-by-step format. However, the major areas include:

• Firstly, evaluating text with Azure Cognitive Language Services
• Secondly, processing and translating speech with Azure Cognitive Speech Services
• Thirdly, processing and classifying images with the Azure cognitive vision services
• Next, processing natural language with Azure Cognitive Language Services
• Then, implementing knowledge mining with Azure Cognitive Search
• Lastly, creating conversational AI solutions

Microsoft AI-102 Exam Details

Microsoft Azure AI-102 exam consists of 40-60 questions that can be in the format like scenario-based single answer questions, multiple-choice questions, arrange in the correct sequence type questions, drag & drop questions, mark review, drag, and drop, etc. However, talking about the passing score, a candidate has to score a minimum of 700 or more. Further, the Microsoft AI-102 exam cost is $165 USD and can be given in only the English language.

Scheduling Exam

Microsoft Azure AI-102 exam measures the ability to perform tasks like planning and managing Azure Cognitive Services solutions with implementing Computer Vision solutions and natural language processing solutions. However, for scheduling the exam, candidates can log in to their Microsoft account and fill in the details. 

Microsoft AI-102 Exam Course Outline

Microsoft provides a course outline for the AI-102 exam covering the major sections to help in better preparation. This include:

Microsoft AI-102 Exam Course Outline was updated on August 2, 2022.

1. Plan and Manage an Azure Cognitive Services Solution

Select the appropriate Cognitive Services resource

• select the appropriate cognitive service for a vision solution (Microsoft Documentation: Microsoft cognitive services technology, What is Computer Vision?, What is the Azure Face service? Azure Media Services Video Indexer)
• selecting the appropriate cognitive service for a language analysis solution (Microsoft Documentation: What are Azure Cognitive Services? What is Language Understanding (LUIS)? What is the Translator service? What is the Text Analytics API?)
• select the appropriate cognitive Service for a decision support solution (Microsoft Documentation: What is the Anomaly Detector API? What is Azure Content Moderator? What are Azure Cognitive Services?)
• select the appropriate cognitive service for a speech solution (Microsoft Documentation: What are Azure Cognitive Services? What is the Speech service? What is speech-to-text? What is text-to-speech?)

Plan and configure security for a Cognitive Services solution

• manage Cognitive Services account keys (Microsoft Documentation: Create a Cognitive Services resource using the Azure portal, az cognitiveservices account keys)
• manage authentication for a resource (Microsoft Documentation: Authenticate requests to Azure Cognitive Services)
• secure Cognitive Services by using Azure Virtual Network (Microsoft Documentation: Configure Azure Cognitive Services virtual networks)
• plan for a solution that meets responsible AI principles (Microsoft Documentation: Identify guiding principles for responsible AI)

Create a Cognitive Services resource

• create a Cognitive Services resource (Microsoft Documentation: Create a Cognitive Services resource using the Azure Command-Line Interface(CLI), Using the Azure portal)
• configure diagnostic logging for a Cognitive Services resource (Microsoft Documentation: Enable diagnostic logging for Azure Cognitive Services)
• manage Cognitive Services costs (Microsoft Documentation: Plan and manage costs for Azure Cognitive Services)
• monitor a cognitive service (Microsoft Documentation: Monitor operations and activity of Azure Cognitive Search)
• implement a privacy policy in Cognitive Services (Microsoft Documentation: Data, privacy, and security for Spatial Analysis)

Plan and implement Cognitive Services containers

• identify when to deploy to a container (Microsoft Documentation: Azure Cognitive Services containers)
• containerize Cognitive Services (including Computer Vision API, Face API, Text Analytics, Speech, Form Recognizer) (Microsoft Documentation: Install and run Docker containers for the Anomaly Detector API, Install and run Docker containers for LUIS, Speech service APIs, Text Analytics containers, Face containers)
• Deploy Cognitive Services containers in Microsoft Azure (Microsoft Documentation: Azure Cognitive Services containers)

2. Implement Computer Vision Solutions

Analyze images by using the Computer Vision API

• retrieve image descriptions and tags by using the Computer Vision API (Microsoft Documentation: Describe images with a human-readable language, Applying content tags to images)
• identify landmarks and celebrities by using the Computer Vision API (Microsoft Documentation: Detect domain-specific content)
• detect brands in images by using the Computer Vision API (Microsoft Documentation: Detect popular brands in images)
• moderate content in images by using the Computer Vision API (Microsoft Documentation: Learn image moderation concepts)
• generate thumbnails by using the Computer Vision API (Microsoft Documentation: Generating smart-cropped thumbnails with Computer Vision)

Extract text from images

• extracting text from images or PDFs by using the Computer Vision service (Microsoft Documentation: Optical character recognition)
• extract information using the pre-built receipt model in Form Recognizer (Microsoft Documentation: Form Recognizer prebuilt receipt model)
• build and optimize a custom model for Form Recognizer (Microsoft Documentation: Use the Form Recognizer client library or REST API, Build a training data set for a custom model)

Extract facial information from images

• detect faces in an image by using the Face API (Microsoft Documentation: Get face detection data)
• recognize faces in an image by using the Face API (Microsoft Documentation: Add faces to a PersonGroup, Use the Face client library)
• match similar faces by using the Face API (Microsoft Documentation: Face – Find Similar)

Implement image classification by using the Custom Vision service

• label images by using the Custom Vision Portal (Microsoft Documentation: Label images faster with Smart Labeler)
• train a custom image classification model in the Custom Vision Portal (Microsoft Documentation: Build a classifier with the Custom Vision website)
• train a custom image classification model by using the SDK (Microsoft Documentation: Create an image classification project with the Custom Vision client library or REST API)
• manage model iterations (Microsoft Documentation: Build a classifier with the Custom Vision website, Use your model with the prediction API)
• evaluate classification model metrics (Microsoft Documentation: Build a classifier with the Custom Vision website)
• publish a trained iteration of a model (Microsoft Documentation: Use your model with the prediction API)
• export a model in an appropriate format for a specific target (Microsoft Documentation: Export your model for use with mobile devices)
• consume a classification model from a client application (Microsoft Documentation: Consume an Azure Machine Learning model deployed as a web service)
• deploy image classification custom models to containers (Microsoft Documentation: Perform image classification at the edge with Custom Vision Service)

Implement an object detection solution by using the Custom Vision service

• label images with bounding boxes by using the Custom Vision Portal (Microsoft Documentation: Tag images in a labeling project)
• train a custom object detection model by using the Custom Vision Portal (Microsoft Documentation: Build an object detector with the Custom Vision website)
• train a custom object detection model by using the SDK (Microsoft Documentation: Create an object detection project with the Custom Vision client library)
• manage model iterations (Microsoft Documentation: Manage training iterations)
• evaluate object detection model metrics (Microsoft Documentation: Evaluate the detector)
• publish a trained iteration of a model (Microsoft Documentation: Publish the current iteration)
• consume an object detection model from a client application (Microsoft Documentation: Use the object detection model in Power Automate)
• deploy custom object detection models to containers (Microsoft Documentation: Develop Custom Object Detection Models with NVIDIA and Azure Machine Learning)

Analyze video by using Azure Video Analyzer for Media (formerly Video Indexer)

• process a video (Microsoft Documentation: Upload and index your videos)
• extract insights from a video (Microsoft Documentation: Analyze video)
• moderate content in a video (Microsoft Documentation: Azure Content Moderator)
• customize the Brands model used by Video Indexer (Microsoft Documentation: Customize a Brands model with the Video Indexer website)
• customize the Language model used by Video Indexer by using the Custom Speech Service (Microsoft Documentation: Customize a Language model with the Video Indexer website)
• customize the Person model used by Video Indexer (Microsoft Documentation: Customize a Person model with the Video Indexer website)
• extract insights from a live stream of video data (Microsoft Documentation: Live stream analysis with Video Indexer)

3. Implement Natural Language Processing Solutions

Analyze text by using the Language service

• retrieve and process key phrases (Microsoft Documentation: extract key phrases using Text Analytics)
• retrieving and processing entity information (people, places, URLs, etc.) (Microsoft Documentation: Supported entity categories in the Text Analytics API v3, How to use Named Entity Recognition in Text Analytics)
• retrieve and process sentiment (Microsoft Documentation: Sentiment analysis on streaming data using Azure Databricks, Build a Flask app with Azure Cognitive Services)
• detect the language used in the text (Microsoft Documentation: Detect language with Text Analytics)

Manage speech by using the Speech service

• implement text-to-speech (Microsoft Documentation: text-to-speech)
• customize text-to-speech (Microsoft Documentation: Custom Voice)
• implement speech-to-text (Microsoft Documentation: speech-to-text)
• improve speech-to-text accuracy (Microsoft Documentation: Create a tenant model)
• Improve text-to-speech accuracy (Microsoft Documentation: Improve recognition accuracy with phrase list)
• Implement intent recognition (Microsoft Documentation: Recognize intents with the Speech service and LUIS)

Translate language

• translate text by using the Translator service (Microsoft Documentation: Create a translation app with WPF)
• translating speech-to-speech by using the Speech service (Microsoft Documentation: speech translation)
• translate speech-to-text by using the Speech service (Microsoft Documentation: speech-to-text)

Build an initial language model by using Language Understanding

• create intents and entities based on a schema, and add utterances (Microsoft Documentation: Add entities to extract data, Add intents to determine user intention of utterances, Extract data with entities, Intents in your LUIS app)
• create complex hierarchical entities (Microsoft Documentation: Entity types)
• train and deploy a model (Microsoft Documentation: Deploy an app in the LUIS portal)

Iterate on and optimize a language model by using language understanding

• implement phrase lists (Microsoft Documentation: Create a phrase list for a concept, Add phrase list as a feature)
• implement a model as a feature (i.e. prebuilt entities) (Microsoft Documentation: Use features to boost signal of word list, Extract data with entities, Add a prebuilt entity)
• manage punctuation and diacritics (Microsoft Documentation: Punctuation normalization, Diacritics, Diacritics normalization)
• implement active learning (Microsoft Documentation: Concepts for enabling active learning by reviewing endpoint utterances, Log user queries to enable active learning)
• monitor and correct data imbalances (Microsoft Documentation: Review data imbalance)
• implement patterns (Microsoft Documentation: Patterns improve prediction accuracy, Add common pattern template utterance formats to improve predictions)

Manage a language understanding model

• manage collaborators (Microsoft Documentation: Add contributors to your app)
• managing versioning (Microsoft Documentation: Use versions to edit and test without impacting staging or production apps, Application, and version settings)
• publish a model through the portal or in a container (Microsoft Documentation: Publish your active, trained app to a staging or production endpoint)
• export a Language Service package (Microsoft Documentation: Export packaged app from LUIS, Export and delete your customer data in Language Understanding (LUIS) in Cognitive Services)
• deploy a Language Service package to a container (Microsoft Documentation: Deploy and run a container on Azure Container Instance)

Create a Questions Answering solution using the Language service

• Create a question answering project (Microsoft Documentation: Create, test, and deploy a custom question answering project)
• Import questions and answers (Microsoft Documentation: Format guidelines for question answering)
• train and test a knowledge base (Microsoft Documentation: Test your knowledge base in QnA Maker, Create, train, and publish your QnA Maker knowledge base)
• publish a knowledge base (Microsoft Documentation: Publish the knowledge base)
• create a multi-turn conversation (Microsoft Documentation: create multiple turns of a conversation)
• add alternate phrasing (Microsoft Documentation: Add additional alternatively-phrased questions, Add alternate questions)
• add chit-chat to a knowledge base (Microsoft Documentation: Add Chit-chat to a knowledge base)
• export a knowledge base (Microsoft Documentation: Migrate a knowledge base using export-import)
• add active learning to a knowledge base (Microsoft Documentation: Active learning)

4. Implement Knowledge Mining Solutions

Implement a Cognitive Search solution

• create data sources (Microsoft Documentation: Create Data Source)
• define an index (Microsoft Documentation: Creating search indexes in Azure Cognitive Search)
• create and run an indexer (Microsoft Documentation: Creating indexers in Azure Cognitive Search)
• query an index (Microsoft Documentation: Querying in Azure Cognitive Search)
• configure an index to support autocomplete and autosuggest (Microsoft Documentation: Add autocomplete and suggestions to client apps using Azure Cognitive Search, Create a suggester to enable autocomplete and suggested results in a query)
• boost results based on relevance (Microsoft Documentation: Add scoring profiles to an Azure Cognitive Search index)
• implement synonyms (Microsoft Documentation: Synonyms in Azure Cognitive Search)

Implement an AI enrichment pipeline

• attach a Cognitive Services account to a skillset (Microsoft Documentation: Attach a Cognitive Services resource to a skillset in Azure Cognitive Search)
• select and include built-in skills for documents (Microsoft Documentation: Built-in cognitive skills for text and image processing during indexing, Document Extraction cognitive skill)
• implement custom skills and include them in a skillset (Microsoft Documentation: add a custom skill to an Azure Cognitive Search enrichment pipeline)

Implement a knowledge store

• define file projections (Microsoft Documentation: Projecting to file)
• defining object projections (Microsoft Documentation: Projecting to objects)
• define table projections (Microsoft Documentation: Projecting to tables)
• query projections (Microsoft Documentation: Knowledge store “projections” in Azure Cognitive Search)

Manage a Cognitive Search solution

• provision Cognitive Search (Microsoft Documentation: Create an Azure Cognitive Search service in the portal)
• configure security for Cognitive Search (Microsoft Documentation: Security overview for Azure Cognitive Search, Configure customer-managed keys for data encryption in Azure Cognitive Search, Configure IP firewall for Azure Cognitive Search)
• configure scalability for Cognitive Search (Microsoft Documentation: Scale for performance on Azure Cognitive Search)

Manage indexing

• manage re-indexing (Microsoft Documentation: Update Index (Azure Cognitive Search REST API))
• rebuild indexes (Microsoft Documentation: How to rebuild an index in Azure Cognitive Search)
• schedule indexing (Microsoft Documentation: How to schedule indexers in Azure Cognitive Search)
• monitor indexing (Microsoft Documentation: How to monitor Azure Cognitive Search indexer status and results)
• implement incremental indexing (Microsoft Documentation: Incremental enrichment and caching in Azure Cognitive Search)
• manage concurrency (Microsoft Documentation: How to manage concurrency in Azure Cognitive Search)
• push data to an index (Microsoft Documentation: Pushing data to an index)
• troubleshoot indexing for a pipeline (Microsoft Documentation: Indexer troubleshooting guidance for Azure Cognitive Search)

5. Implement Conversational AI Solutions

Design and implement conversation flow

• design conversation logic for a bot (Microsoft Documentation: Design and control conversation flow)
• create and evaluate .chat file conversations by using the Bot Framework Emulator (Microsoft Documentation: Debug your bot using transcript files)
• Choose an appropriate conversational model for a bot, including activity handlers and dialogs (Microsoft Documentation: Dialogs as conversational building blocks in Composer)

Create a bot by using the Bot Framework SDK

• Use the Bot Framework SDK to create a bot from a template (Microsoft Documentation: Create a bot with the Bot Framework SDK)
• Implement activity handlers and dialogs (Microsoft Documentation: Event-driven conversations using an activity handler, ActivityHandler class)
• Use a turn context (Microsoft Documentation: TurnContext class)
• Test a bot using the Bot Framework Emulator (Microsoft Documentation: Test and debug with the Emulator)
• Deploy a bot to Azure (Microsoft Documentation: Provision and publish a bot)

Create a bot by using the Bot Framework Composer

• implement dialogs (Microsoft Documentation: Dialogs Overview)
• maintain state (Microsoft Documentation: Managing state)
• implement logging for a bot conversation (Microsoft Documentation: Add telemetry to your bot)
• implement prompts for user input (Microsoft Documentation: Ask for user input)
• troubleshoot a conversational bot (Microsoft Documentation: General troubleshooting for Azure Bot Service bots)
• test a bot (Microsoft Documentation: Test and debug with the Emulator)
• publish a bot (Microsoft Documentation: Publish a bot)
• Add language generation for a response (Microsoft Documentation: Language generation in Composer)
• Design and implement Adaptive Cards (Microsoft Documentation: Adaptive Cards Overview)

Integrate Cognitive Services into a bot

• integrating a question answering model (Microsoft Documentation: question answering)
• integrate a language understanding service (Microsoft Documentation: Add natural language understanding to your bot)
• integrating a Speech service (Microsoft Documentation: Add speech to messages with the Bot Connector API)

For Queries: Check Microsoft Azure AI-102 Exam FAQs

Exam Policies

Microsoft Certification exam policies provide all the exam-related details and information with exam giving procedures. These exam policies have certain rules that need to be followed during the exam time or at testing centers. Some of them include:

Exam retake policy

This states that candidate’s failing the exam for the first time must wait 24 hours before retaking the exam. During this time, they can go onto the certification dashboard and reschedule the exam. If this happens for the second time then, they may have to wait for at least 14 days before retaking the exam. Likewise, this 14-day waiting period is also imposed between the third and fourth attempts and fourth and fifth attempts. However, candidates can only give any exam five times a year. And, the 12 month period starts from the very first attempt.

Exam reschedule and the cancellation policy

Microsoft will be temporarily waiving the reschedule and cancellation fee if candidates cancel their exams within 24 hours before the scheduled appointment. However, for rescheduling or canceling an appointment no charge will be applied at least 6 business days prior to your appointment. But, if a candidates cancel or reschedule an exam within 5 business days of your registered exam time then, a fee will be applied. Lastly, if a candidate failed to show up for an exam appointment or forgot to reschedule an appointment at least 24 hours then they can forfeit your entire exam fee.

Microsoft Azure AI-102 Exam Study Guide

Exam objectives

For having a good start for the Microsoft AI-102 exam preparation, candidates must have familiarity with the exam objectives. The Microsoft AI-102 exam objectives cover four important topics that will provide understanding in the major sections. So, check out the exam guide that covers the following topics:

• Firstly, plan and manage an Azure Cognitive Services solution
• Secondly, implement Computer Vision solutions
• Implement natural language processing solutions
• Then, implement knowledge mining solutions
• Lastly, implement conversational AI solutions

Microsoft Learning Platform

Microsoft gives access to learning platforms covering various resources that will help in exam preparation. For the AI-102 exam preparation, go through the Microsoft official website to get all the necessary information about the AI-102 exam. Furthermore, this also provides benefits to the candidates to understand the concepts and to pass the exam.

Microsoft Docs

Microsoft documentation is the knowledge source that provides detailed information about the AI-102 exam concepts. Moreover, using Microsoft Documentation, you will get to know the different scales of different Azure AI services. This consists of modules that will help you gain a lot of knowledge about the different AI services and concepts used according to the exam.

Instructor-led Training

Designing and Implementing an Azure AI Solution (AI-102) exam suitable for software developers who want to build AI-infused applications that supports Azure Cognitive Services, Azure Cognitive Search, and Microsoft Bot Framework. However, this Microsoft AI-102 training course will use C# or Python as the programming language.

Target Audience

• Firstly, those having knowledge and skills to build, manage and deploy AI solutions that support Azure Cognitive Services, Azure Cognitive Search, and Microsoft Bot Framework. 
• Secondly, those with having familiarity with C# or Python.
• Lastly, candidates with knowledge of using REST-based APIs for building computer vision, language analysis, knowledge mining, intelligent search, and conversational AI solutions on Azure.

Online Study Groups

Online study groups can provide benefits to candidates during exam preparation. That is to say, joining the study groups will help you to stay connected with the experts and professionals who are already on this pathway. Moreover, you can start discussing your query or the issue related to the exam in this group and take the AI-102 exam study notes.

Practice Tests

Practice tests are important for having better preparation. By assessing yourself with Microsoft AI-102 practice tests, you will know about your weak and strong areas. Moreover, you will be able to improve your answering skills that as a result will save a lot of time during the exam. The smart way to take the AI-102 exam practice tests is after completing a full topic and then try the mock tests. This will also make your revision strong. So, find the best practice exam tests and get yourself prepared for the Microsoft AI-102 certification exam.

Start preparing for Microsoft AI-102 Exam Now!