Questions Flashcards

(50 cards)

1
Q
  1. Our organization has hundreds of IoT devices with sensors. We want to ensure that we can write-back configuration changes to the devices. Which would we use?

A. Microsoft Azure Event Hubs

B. Microsoft Azure Notification Hubs

C. Microsoft Azure IoT Hub

D. API apps in Microsoft Azure App Service

A

Correct Answer: C

Why is this correct?
IoT Hub is the best choice for sending data from IoT devices as it provides device-to-cloud messaging for all supported IoT devices.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q
  1. We are designing an AI workflow solution that needs to use data that must be sourced from an on-premises Oracle database. What should we use to ingest the data?

A. Azure SQL Data Warehouse

B. Azure Databricks

C. Azure Data Lake Storage

D. Azure Data Factory

A

Correct Answer: D

Why is this correct?
Azure Data Factory can be used to source data from an Oracle database.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q
  1. We are designing an AI solution that will need to analyse millions of images. We want to minimize storage costs. Which storage solution should we use?

A. Azure Data Lake Storage Gen2

B. Azure Table storage

C. Azure Data Lake Storage Gen1

D. Azure File storage

A

Correct Answer: A
Why is this correct?
Azure Data Lake Storage Gen2 is the most cost-effective form of storage for this quantity of images. Azure Data Lake Storage Gen1 is possible, but Gen2 is a better, newer, and more cost-effective solution.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q
  1. We are designing an AI solution that will replace an existing system utilizing data in a SQL Server database. Which of the following are the simplest means of cloud-enabling the SQL Server data in this scenario?

A. Azure SQL Database

B. Azure Cosmos DB

C. Azure Databricks

D. Azure Table storage

A

Correct Answer: A

Why is this correct?
Azure SQL Database is the simplest solution as it is possible we can simply migrate the schema and data, then change connection strings in the application.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q
  1. We would like to moderate a set of data consisting of text and images for offensive content. Which two services should we use?

A. Term List API

B. Text Moderation API

C. Face API

D. Image Moderation API

A

Correct Answer: B
Why is this correct?
This would perform the text portion of the moderation.

Correct Answer: D
Why is this correct?
This would perform the image portion of the moderation.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q
  1. We would like to collect temperatures from IoT devices and take action on the average temperature over every one-minute interval. We want to minimize the amount of uploaded data. Where would we run the solution?

A. Azure Event Hubs

B. An Azure IoT Edge device

C. A WebJob in Azure

D. Azure Notification Hubs

A

Correct Answer: B

Why is this correct?
By running on the IoT Edge device, we can aggregate the data before sending it to Azure, hence minimizing the amount of data uploaded.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q
  1. We are designing an AI workflow solution that needs to use data sourced from an on-premises SQL Server database. What should we use to ingest the data?

A. Azure SQL Data Warehouse

B. Azure Data Lake Storage

C. Azure Data Factory

D. Azure Databricks

A

Correct Answer: C

Why is this correct?
Azure Data Factory can be used to source data from an on-premises SQL Server database.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q
  1. We are designing an AI workflow solution that needs to use data sourced from JSON files stored in an Azure Storage account. What should we use to ingest the data?

A. Azure Databricks

B. Azure Data Factory

C. Azure SQL Data Warehouse

D. Azure Data Lake Storage

A

Correct Answer: B

Why is this correct?
Azure Data Factory can be used to integrate JSON data in an Azure Storage account.

Azure Data Lake Storage is a data store for big data analytics, not a means of accessing or ingesting JSON data.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q
  1. We need to build a solution to monitor the sentiment of tweets in a Twitter feed. When negative tweets are identified, we want to send an email. We want to minimize the amount of custom code. What should we use?

A. Azure Functions

B. Azure Stream Analytics

C. Azure Databricks

D. Azure Logic Apps

E. Azure Cognitive Services

A

Correct Answer: D

Why is this correct?
Logic Apps minimize code and can be used to implement this solution.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q
  1. We want to build a sentiment analysis solution the can use PDF data as input and must process the data in batches. Which should we use to store the documents?

A. Azure Managed Disks

B. Azure Blob storage

C. Azure Table storage

D. Azure Cosmos DB

A

Correct Answer: B

Why is this correct?
Blobs are the best solution for PDF as Cognitive Services text analysis, and Azure Cognitive Search can examine the PDF content in the blobs.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q
  1. We are designing an AI workflow solution that needs to use data that needs to be sourced from an Azure SQL database. What should we use to ingest the data?

A. Azure Data Factory

B. Azure Databricks

C. Azure Data Lake Storage

D. Azure SQL Data Warehouse

A

Correct Answer: A

Why is this correct?
Azure Data Factory can be used to source data from an Azure SQL database.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q
  1. Our organization has hundreds of IoT devices with sensors. We want to process the sensor data in Azure. Which messaging service should we use to send the messages to Azure?

A. Microsoft Azure Event Hubs

B. API apps in Microsoft Azure App Service

C. Microsoft Azure Notification Hubs

D. Microsoft Azure IoT Hub

A

Correct Answer: D

Why is this correct?
IoT Hub is the best choice for sending data from IoT devices as it provides device-to-cloud messaging for all supported IoT devices.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q
  1. We want to build a sentiment analysis solution that uses JSON data as input and must process the data in batches. Which should we use to store the documents?

A. Azure Table storage

B. Azure Blob storage

C. Azure Managed Disks

D. Azure Cosmos DB

A

Correct Answer: D

Why is this correct?
Cosmos DB is a great solution for native storage, retrieval, and processing of JSON documents

Blobs could work, but there is a better solution for JSON and integration with Cognitive Services text analytics.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q
  1. We plan to store over 10,000 images of approximately 5MB on IoT Edge devices for up to 14 days. What should we use?

A. Azure Blob storage on the IoT Edge devices

B. Azure Database for Postgres SQL.

C. Microsoft SQL Server on the IoT Edge devices

D. Azure Stream Analytics on the IoT Edge devices.

A

Correct Answer: A

Why is this correct?
Blob storage can be run on the IoT device and can provide this capability.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q
  1. We want to build a solution that will tag scenes in a video and then be able to search by tags to locate matching scenes. Which Cognitive Services API would we use?

A. Custom Vision

B. Computer Vision

C. Bing Video Search

D. Video Indexer

A

Correct Answer: D

Why is this correct?
Azure Video Indexer can be used to index videos and add metadata for search.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q
  1. Which type of bot template would we use for a bot that proactively sends messages to a user, such as upon the run of an Azure Function or on a timer?

A. Q&A

B. Basic

C. Proactive

D. Form

E. Language Understanding

A

Correct Answer: C

Why is this correct?
This is a template used to create a bot that can send proactive messages to the user. These proactive scenarios can be tied to respond to messages in queues, trigger functions based on queue messages, and interoperate with bots

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q
  1. We plan to make a bot that will answer common user questions. The bot needs to support two languages and be used in three different regions. The bot will use LUIS to identify which language is being used. How many instances of QnA Maker will we need to deploy?

A. 2

B. 3

C. 4

D. 1

A

Correct Answer: D

Why is this correct?
QnA maker can support multiple languages in a single instance, so only one is required.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q
  1. We are designing a solution to analyse data with PII, and we want to make sure the data is not stored or processed in the cloud. What is our recommendation?

A. Azure Machine Learning Studio.

B. Azure Machine Learning services.

C. A Text Analytics API that has container support.

D. An Apache Spark cluster that uses MMLSpark.

A

Correct Answer: C

Why is this correct?
Text Analytics run in a container outside of the cloud can be used to keep the data stored and processed on devices not in the cloud. Install and run Text Analytics containers

While Apache Spark is awesome technology, it isn’t the best solution for this scenario as its an analytics service and not best suited to this requirement. As a matter of fact, it would require storing data in the cloud.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q
  1. Which type of bot card can be used to present a receipt to the user, often containing a list of items, taxes, totals, and other text?

A. Audio card

B. Animation card

C. Thumbnail card

D. Video card

E. Receipt card

F. Adaptive card

G. Sign-in card

H. Hero card

A

Correct Answer: E

Why is this correct?
This is a card that enables a bot to provide a receipt to the user. It typically contains the list of items to include on the receipt, taxes, totals, and other text.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q
  1. Our company plans to deploy an app to identify named entities in text. Which Cognitive Services API would we use?

A. Text Analytics

B. Language Understanding (LUIS)

C. Bing Visual Search API

D. Content Moderator

A

Correct Answer: A

Why is this correct?
This is one of the capabilities of Text Analytics.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q
  1. Which Bot Framework toolkit would we use to prototype mock conversations?

A. LUIS

B. Dispatch

C. Chatdown

D. LUDown

E. QnA Maker

A

Correct Answer: C

Why is this correct?
Chatdown can be used to prototype mock conversations in markdown and convert the markdown to transcripts we can load and view in the new V4 Bot Framework Emulator.

22
Q
  1. We need to create a bot that users can use after signing in using a Gmail user account. In which order should we complete the following tasks?

A. From the Azure portal, configure MFA for AAD. From the Azure Portal, create an Azure Active Directory (Azure AD) B2C service. From the Azure Portal, create a client application. From the bot code, add the connection settings and OAuthPrompt.

B. Create a client application in AAD. From the Azure Portal, create an Azure Active Directory (Azure AD) B2C service. From the Azure Portal, configure an identity provider. From the bot code, add the connection settings and OAuthPrompt.

C. From the bot code, add the connection settings and OauthPrompt. From the Azure Portal, create a client application. From the Azure Portal, configure an identity provider. From the Azure Portal, create an Azure Active Directory (Azure AD) B2C service.

D. From the Azure Portal, configure an identity provider. From the Azure Portal, create an Azure Active Directory (Azure AD) B2C service. From the Azure Portal, create a client application. From the bot code, add the connection settings and OAuthPrompt.

A

Correct Answer: B

Why is this correct?
This is the correct sequence for configuring required B2C connectivity, and these must be performed before configuring OAuthPrompt in the bot. Set up sign-up and sign-in with a Google account using Azure Active Directory B2C

We need to create the AAD B2C before configuring the identity provider. Set up sign-up and sign-in with a Google account using Azure Active Directory B2C

NOTE: most likely a drag-and-drop in the exam!

23
Q
  1. Which type of bot template would we use for a bot that uses dialogs to respond to user input?

A. Proactive

B. Form

C. Language Understanding

D. Basic

E. Q&A

A

Correct Answer: D

Why is this correct?
This is a template used to create a bot that uses dialogs to respond to user input
Form is a template used to create a bot that collects input from a user via a guided conversation.

24
Q
  1. Which type of bot card would we use to play an audio file?

A. Video card

B. Sign-in card

C. Adaptive card

D. Animation card

E. Receipt card

F. Thumbnail card

G. Hero card

H. Audio card

A

Correct Answer: H

Why is this correct?
This is a card that can play an audio file.

25
25. We are designing an AI solution to monitor meetings and we want to identify attendees based upon faces in video. Which cognitive service should we use? A. Computer Vision B. Face API C. Speech-to-text D. QnA Maker E. Text Analytics
Correct Answer: B Why is this correct? The Face API can be used to identify specific people in videos and images.
26
26. Which Bot Framework toolkit would we use to build language models allowing us to dispatch between disparate components (such as QnA, LUIS, and custom code)? A. Dispatch B. LUDown C. LUIS D. QnA Maker E. Chatdown
Correct Answer: A Why is this correct? Dispatch can be used to build language models allowing us to dispatch between disparate components (such as QnA, LUIS, and custom code).
27
27. Our company plans to deploy an app to interpret user requests sent by users entering text input. Which Cognitive Services API would we use? A. Language Understanding (LUIS) B. Text Analytics C. Bing Visual Search API D. Content Moderator
Correct Answer: A Why is this correct? LUIS is used to determine the intent of the user.
28
28. We need to create an IoT/AI Solution that takes images every minute and counts the number of people in the image at the edge. How would you implement the edge compute? A. A C# app on a Raspberry Pi. B. A Python app on an Arduino. C. A custom app with Windows IoT and OpenCV D. Cognitive Services containers on the devices.
Correct Answer: D Why is this correct? IoT Edge and Cognitive Service containers on edge devices are a simple means of deploying IoT and AI solutions to the edge computing devices
29
29. Which type of bot card would we use to play animated GIFs or videos? A. Animation card B. Audio card C. Adaptive card D. Receipt card E. Hero card F. Thumbnail card G. Video card H. Sign-in card
Correct Answer: A Why is this correct? This is a card that can play animated GIFs or short videos. Video card can play videos.
30
30. Which Bot Framework toolkit would we use to create and manage QnAMaker.ai Knowledge Bases? A. LUDown B. LUIS C. Dispatch D. QnA Maker E. Chatdown
Correct Answer: D Why is this correct? QnA Maker can be used to create and manage QnAMaker.ai Knowledge Bases.
31
31. We'd like to identify and visualize unusual fluctuations in a pricing indicator with data spanning 10 years. Which Cognitive Service would we use? A. Text Analytics B. Bing Autosuggest C. Computer Vision D. Anomaly Detector API
Correct Answer: D Why is this correct? The Anomaly Detector API enables us to monitor and detect abnormalities in our time series data with machine learning.
32
32. Which type of bot card can be used to display a single thumbnail image, one or more buttons, and text? A. Sign-in card B. Hero card C. Receipt card D. Audio card E. Animation card F. Video card G. Thumbnail card H. Adaptive card
Correct Answer: G Why is this correct? This is a card that typically contains a single thumbnail image, one or more buttons, and text. Hero card typically contains a single large image, one or more buttons, and text.
33
33. We have created a cognitive services resource, and a DevOps person needs to retrieve the keys for the service to deploy to a deployment environment. We want to use RBAC to limit access to the least privilege. Which RBAC role would we use to control access to the CS resource by the DevOps person? A. Cognitive Services User B. Security Manager C. Security Reader D. Cognitive Services Contributor
Correct Answer: A Why is this correct? This role allows the user to only list keys in the Cognitive Service.
34
34. We are building an AI solution that will analyze data stored in blobs. We need to encrypt the blob data using customer keys. Which three actions should we take to enable the encryption? A. Enable encryption that uses customer-managed keys. B. Upload the key to an Azure Key Vault. C. Generate a shared access signature (SAS). D. Configure a service endpoint for the storage account. E. Generate an encryption key.
Correct Answer: A Why is this correct? To use customer-managed keys in Key Vault, we must first enable the capability. Correct Answer: E Why is this correct? An encryption key would be created by the customer and provided to Key Vault for storage. Correct Answer: B Why is this correct? Once the user generates a key and customer-managed keys are enabled, the customer-generated key can be stored in Key Vault and used for blob encryption. Service endpoints control access from virtual networks to services such as storage accounts but do not provide the actual encryption of the blob data.
35
35. We plan to build a bot to facilitate with the help desk of our company. The bot must be able to give answers from an existing knowledge base. Which solution should we recommend? A. The Dispatch tool library. B. The QnA Maker service. C. The Speech-to-text API. D. The Text Analytics API E. The Language Understanding (LUIS) service
Correct Answer: B Why is this correct? QnA Maker is used to provide FAQ-like bot capabilities.
36
36. We are designing an AI solution to monitor meetings and want to know when facial expressions indicate people being angry or scared. Which cognitive service should we use? A. Speech-to-text B. QnA Maker C. Face API D. Text Analytics E. Computer Vision
Correct Answer: C Why is this correct? The Face API is capable of identifying facial expressions.
37
37. Which type of bot card can be used to present a receipt to the user, often containing a list of items, taxes, total information, and other text? A. Sign-in card B. Adaptive card C. Animation card D. Receipt card E. Thumbnail card F. Audio card G. Video card H. Hero card
Correct Answer: D Why is this correct? This is a card that enables a bot to provide a receipt to the user. It typically contains the list of items to include on the receipt, tax and total information, and other text.
38
38. We would like to test our production LUIS application locally on Azure Stack outside of Azure. Which three of the following should we perform? A. Get the docker image for LUIS from mcr.microsoft.com/azure-cognitive-services/luis:latest. B. Export the production version of the Language Understanding (LUIS) app. C. Deploy an Azure Kubernetes Service (AKS) cluster into Azure. D. Export the staging version of the Language and Understanding (LUIS) app. E. Get the docker image from mcr.microsoft.com/azure-cognitive-services/sentiment:latest. F. Deploy a Kubernetes cluster to Azure Stack. G. Deploy an Azure Container Service cluster to the West Europe region.
Correct Answer: A Why is this correct? We will need to run LUIS on our Azure Stack or AKS cluster, and we can get the container from this location: Get the container image with docker pull. Correct Answer: B Why is this correct? We can export our LUIS app from Azure to run it elsewhere, such as on Azure Stack. Use versions to edit and test without impacting staging or production apps - Other Actions Correct Answer: F Why is this correct? LUIS apps can be run in Azure Stack by deploying the LUIS components to AKS on Azure Stack. Deploy Kubernetes to use containers with Azure Stack
39
39. We have created a bot using C#. We want to deploy the bot to Azure. Which tool should we use? A. The Git CLI B. The Azure CLI C. The .NET Core CLI D. The AzCopy tool
Correct Answer: B Why is this correct? Azure CLI can be used to deploy a bot to Azure. The .NET Core CLI manages .NET core and does not have the ability to deploy LUIS apps to Azure.
40
40. We are designing an AI solution using drones as IoT Edge devices. We want to stream video as a stream of images from the drones into Azure. What service should we use to deliver the images? A. Azure IoT Edge B. Azure IoT Hub C. Azure Functions D. Azure Logic Apps
Correct Answer: B Why is this correct? IoT Hub is the preferred means of sending data from IoT devices into Azure. Functions could provide compute capabilities on the drone or in Azure, but does not have an inherent ability to send data or messages into Azure.
41
41. Our company has thousands of IoT sensors that we plan to use for an AI solution, and we need to recommend a compute solution for real-time analysis of the generated data. What solution would we recommend? A. An Azure HDInsight Kafka cluster B. Azure Notification Hubs C. An Azure HDInsight Spark cluster D. An Azure HDInsight Hadoop cluster
Correct Answer: A Why is this correct? An Azure HDInsight Kafka cluster is a real-time analytics engine and is perfect for this type of solution. HDInsight Hadoop would be used for managing big-data analytics and not the best solution among these answers.
42
42. We want to deploy and orchestrate our AI app using Kubernetes (AKS). Which three actions should we undertake? A. Create a web app for containers. B. Create an Azure Container Registry instance (ACR). C. Create a Kubernetes cluster. D. Create a container image file and deploy it to Azure Container Registry (ACR). E. Create an Azure Container Instance.
Correct Answer: B Why is this correct? Containers deployed to AKS must be stored in a container registry. Azure provides ACR, which has tight integration with AKS, but we could also use others such as Docker Hub. Container Registry Correct Answer: C Why is this correct? To deploy our AI app to AKS, we first need to create an AKS cluster. Quickstart - Deploy an Azure Kubernetes Service cluster using the Azure CLI Correct Answer: D Why is this correct? Containers deployed to AKS will typically be stored in ACR, although it is possible to use other repositories such as Docker Hub. Authenticate with Azure Container Registry from Azure Kubernetes Service
43
43. We need to process images to extract text content (OCR). Which Cognitive Service would we use? A. Translator Text B. Image Moderation API C. Computer Vision D. Text Analytics
Correct Answer: C Why is this correct? The Computer Vision API provides OCR capabilities, along with many other helpful vision features. What is Computer Vision?
44
44. We are running an ML model in AKS. We want to update the model and enable Application Insights. Which of the following would we use? A. The Azure Portal B. The Azure CLI C. ML.NET Model Builder D. The Azure Machine Learning SDK for Python
Correct Answer: B Why is this correct? The CLI is likely the most efficient means of updating the AKS deployment to utilize Application Insights, particularly by using the appInsightsEnabled parameter. The Azure Machine Learning SDK for Python provides ML capabilities, not deployment of models into AKS.
45
45. We plan to deploy a solution using Text Analytics and Computer Vision services in both the West US and East Europe Azure Regions. What is the minimum number of API keys required? A. 3 B. 2 C. 1 D. 4
Correct Answer: C Why is this correct? We can deploy one Cognitive Service in one of the regions and access it from apps in both. Access to that one endpoint can be via a single key. If multiple endpoints are used, then we need multiple keys. That is not the minimum scenario as the number of keys is equivalent to the number of endpoints, excluding secondary keys. Both Text Analytics and Computer Vision are accessed using the same endpoint, so the required functionality is handled by a single endpoint. However, this may not be the most reliable solution. If the one region loses connectivity to the other region, then this will fail. It would also introduce latency for that region. So while on is the minimum, it is technically better to deploy two endpoints, one in each region. Two service endpoints would require two keys, one for each (taking into account just primary and not secondary keys). This is above the minimum required, as we only need one endpoint to service both regions for all functionality. One endpoint exposed both Text Analytics and Computer Vision. One endpoint is not optimal as there could be cross-region connectivity issues, but it is the minimum.
46
46. We've provisioned an Azure Search service and need to deploy Cognitive Search. What should we do next? A. Search by using the REST API. B. Search by using the .NET SDK. C. Load data. D. Create an index.
Correct Answer: D Why is this correct? The next step, even before loading data, is to define an index that tells the search service how to parse and index the data
47
47. We have exposed our machine learning model as a web service. We want others to access and be authenticated to the web service. What two things should we provide when calling the web service? A. A resource group name B. Access key C. Endpoint URL D. Subscription ID
Correct Answer: B Why is this correct? An access key authorizes a caller of a machine learning web service and is required to make the call. Correct Answer: C Why is this correct? An endpoint URL for the web service would need to be obtained to to call the web service API.
48
48. We are querying a product catalog using Bing Search service and need to identify the locale of the query, the last 25 query strings, the number of calls to the service, and the top geographic regions. What should we implement? A. Azure API Management (APIM) B. Azure Application Insights C. Bing Statistics D. Azure Monitor
Correct Answer: C Why is this correct? Bing Statistics provides analytics for the Bing Search APIs. These analytics include call volume, top query strings, geographic distribution, and more. We can enable Bing Statistics in the Azure Portal by navigating to our Azure resource and clicking Enable Bing Statistics.
49
49. We are analyzing temperature data on IoT devices and want to send an alert into Azure when the average temperature over a 5 minute period exceeds a threshold. We would like to minimize the amount of code written. What should we use? A. A machine learning model deployed as a web service. B. Azure Functions as an IoT Edge module. C. An Azure Machine Learning model as an IoT Edge module. D. Azure Stream Analytics as an IoT Edge module.
Correct Answer: D Why is this correct? Using Stream Analytics on the edge can provide us with an efficient means of creating a solution to average values in a stream of data based upon time windows
50
50. We need to create an IoT solution that takes images every minute and sends the images to Azure for processing by Cognitive Services. How would we implement this capability? A. Send the images to an IoT Hub, trigger an Azure Function on each IoT Hub message, and use the function to send the data to Cognitive Services. B. Write the image to a blog in Azure Storage, trigger an Azure Function on each blob’s creation, and use the function to send the data to Cognitive Services. C. Send the images to an IoT Hub, configure Azure Automation to process each IoT Hub message, and send the data to Cognitive Services. D. Write the image to a blob in device storage, trigger an Azure Function in Azure on each blob’s creation, and use the function to send the data to Cognitive Services.
Correct Answer: A Why is this correct? IoT Hub would be used to send the data from an IoT device into Azure, where it can trigger an Azure Function, which then calls the Cognitive Service. ‘Write the image to a blog in Azure Storage, trigger an Azure Function on each blob’s creation, and use the function to send the data to Cognitive Services.’ will likely not work as an IoT device is typically not able to write to Blob storage in Azure. Note: Most likely a drag-and-drop in the exam!!