Practice Questions - Amazon AWS Certified Alexa Skill Builder - Specialty

Question 1

An Alexa Skill Builder needs to display an image and some additional text to users of a skill. Which approach will work with any Amazon Alexa enabled device?
A. Use an Alexa display template, such as BodyTemplate1.
B. Send a Simple card to the user.
C. Send a Standard card to the user.
D. Send a LinkAccount card to the user.

Answer 1

C

Display templates (A) only work on devices with screens. Simple cards (B) do not support images. LinkAccount cards (D) are used for account linking, not displaying images and text. Standard cards (C) support both text and images and are therefore compatible with all Alexa-enabled devices, including those without screens (the text will be read aloud).

Question 2

An Alexa Skill Builder receives feedback from users that a specific utterance causes Amazon Alexa to trigger the skill’s AMAZON.HelpIntent rather than triggering the correct intent and slot. How can the Builder reproduce this behavior to troubleshoot the problem?
A. Use the AWS Lambda test feature to send a request with the given intent and slot combination.
B. Set up a unit test in the code base to simulate what happens when the given intent and slot combination are dispatched within the skill.
C. Use the Manual JSON tab on the Test page of the developer console to see what happens when a request for the given intent and slot combination is sent to the skill.
D. Use the Alexa Simulator tab on the Test page of the developer console to test the utterances the users have reported.

Answer 2

D

The discussion and linked blog post strongly suggest that using the Alexa Simulator (option D) is the best approach. The simulator allows testing of live skills and reproducing user-reported issues using their actual utterances. Options A, B, and C are less effective because they don’t directly simulate the user interaction with Alexa, which is the core problem. Option A tests the Lambda function in isolation, Option B tests the code without the Alexa interaction layer, and Option C uses manual JSON, requiring the builder to already know the correct JSON structure which is what they are trying to debug.

Question 3

An Alexa Skill Builder wants to add a short audio clip to a skill with SSML. The Builder records a 6-second audio clip using a high-end microphone and recording application and exports the file with a bitrate of 48 kbps. The file is then uploaded to a public Amazon S3 bucket. The Builder then updates the SSML to return: Image When testing the skill, Amazon Alexa responds with “There was a problem with the requested skill’s response” every time. How should the Builder resolve this error?
A. Record a shorter audio clip.
B. Re-export the audio file with a smaller bitrate.
C. Re-export the audio file to .mp3 instead of .wav
D. Update the interaction model to enable the AudioPlayer interface

Answer 3

C

The correct answer is C because the provided SSML code uses <audio> tag which, according to Amazon Alexa documentation, only supports the MP3 format. Options A and B are incorrect because the length (6 seconds) and bitrate (48 kbps) are within the acceptable parameters for Alexa audio playback. Option D is incorrect because the AudioPlayer interface is likely already enabled if the skill is attempting to play audio; the issue is the unsupported file format (.wav).

Question 4

An Alexa Skill Builder has published a skill to the Alexa Skills Store, but soon after realizes that Amazon Alexa is saying an incorrect word in one of the responses. What is the recommended method for the Builder to correct this issue?

A. Correct the response in a new endpoint, and after updating the In Development version of the skill, redeploy it.
B. Access the In Development version of the skill, alter the intent that was causing the incorrect word, then resubmit the skill for certification.
C. Edit the response in a new endpoint, and after updating the In Development version of the skill, resubmit it for certification.
D. Correct the response in the existing endpoint of the Live version of the skill, then redeploy it.

Answer 4

B

The discussion indicates that option B is the correct answer. Correcting the issue in the “In Development” version allows for testing and verification before re-submission for certification, ensuring the fix is implemented correctly without impacting the live skill’s functionality. Option A is incorrect because while creating a new endpoint might be a valid approach in other scenarios, it’s not the most efficient or recommended method for a simple word correction. Option C is incorrect because while it updates the “In Development” version, it unnecessarily resubmits the entire skill for certification instead of simply updating the live version after thorough testing. Option D is incorrect because directly editing the live version is risky; it could introduce further errors without proper testing and might violate Amazon’s certification process.

Question 5

An Alexa Skill Builder wants to name a skill using a company’s branded acronym, “NAT Systems.” Which invocation name is valid?
A. n a t systems
B. n-a-t systems
C. n. a. t. systems
D. NAT Systems

Answer 5

C. n. a. t. systems

The discussion indicates that the correct answer is C because Amazon’s guidelines for invocation names require lowercase letters and the use of periods to separate the letters of an acronym. Option A uses spaces which are not allowed. Option B uses hyphens which are also not permitted. Option D uses uppercase letters which are also not allowed. Only Option C adheres to the specified constraints.

Question 6

An Alexa Skill Builder is creating a skill that requires the user to authenticate by speaking a PIN before an order status can be retrieved. According to best practices, how should the PIN value be collected?
A. Use AMAZON.SearchQuery
B. Use AMAZON.FOUR_DIGIT_NUMBER
C. Use multiple slots of type AMAZON.NUMBER
D. Use a custom slot with zero to nine as values.

Answer 6

B. Use AMAZON.FOUR_DIGIT_NUMBER

The best answer is B because AMAZON.FOUR_DIGIT_NUMBER is a built-in intent specifically designed for collecting four-digit numbers, which is the standard length for PINs. This ensures a consistent and user-friendly experience.

Option A is incorrect because AMAZON.SearchQuery is intended for search queries, not PIN entry. Option C is incorrect because while it could technically work, it’s less efficient and user-friendly than using a built-in intent designed specifically for four-digit numbers. It also doesn’t explicitly enforce a four-digit length. Option D is incorrect as it would require significantly more effort for the developer to handle potential errors and doesn’t utilize the built-in functionality for this type of data input. It introduces unnecessary complexity where a standard built-in intent already exists.

Question 7

An Alexa Skill Builder published a skill that streams sounds to help users relax. The skill is becoming very popular and is available in many different locales around the world. More than 20,000 users are using the skill every week, and more than 500 users are added daily. The sound files are stored in Amazon S3. What can the Builder do to ensure low latency and the best possible streaming performance?
A. Use Amazon CloudFront to deliver content and cache the audio files across different geographical regions.
B. Migrate the sound files from Amazon S3 into Amazon Redshift to speed up the read operations.
C. Use Elastic Load Balancing to handle the traffic generated by the increasing number of users.
D. Enable cross-region replication on the Amazon S3 bucket policy to improve the skill’s performance.

Answer 7

A

CloudFront is a content delivery network (CDN) that caches content closer to users geographically. This significantly reduces latency for users accessing audio files, leading to better streaming performance.

Option B is incorrect because Amazon Redshift is a data warehouse service, not suitable for serving audio files. Option C is incorrect because Elastic Load Balancing manages traffic to servers, not the delivery of content from S3. While cross-region replication in S3 (Option D) increases availability, it doesn’t directly address low latency for streaming audio; users will still need to connect to a potentially distant S3 region. CloudFront is the best solution for low latency streaming.

Question 8

An Alexa Skill Builder is developing a skill using AWS Lambda. The Builder made some backend code changes, then tested the skill on an Amazon Echo device. When invoking the skill, Amazon Alexa replies with “There was a problem with the requested skill’s response”. How can the Builder troubleshoot this problem?

A. Use Amazon CloudWatch to check the most recent execution log and see if an error is present.
B. Use Amazon DynamoDB and export a copy of the log database, then search for error messages.
C. Use the developer console to rebuild the model, then invoke the skill again with the invocation name.
D. Use the developer console to add AMAZON.LaunchRequest to the interaction model, rebuild the model, then invoke the skill again.

Answer 8

A

The correct answer is A because AWS Lambda functions log their execution details to Amazon CloudWatch. The error message “There was a problem with the requested skill’s response” is a generic error indicating a problem within the Lambda function itself. Checking the CloudWatch logs will reveal the specific error that occurred during the function’s execution, allowing the builder to pinpoint and fix the issue.

Option B is incorrect because DynamoDB is a NoSQL database and is not used for storing Lambda function logs. Option C is incorrect because rebuilding the model is unnecessary if the problem lies within the Lambda function’s code. Option D is incorrect because adding AMAZON.LaunchRequest addresses issues with the skill’s invocation, not the backend processing within the Lambda function.

Question 9

While developing a skill, an Alexa Skill Builder finds that the voice response is too quick and needs to be slowed down. How can the Builder MOST efficiently iterate and test how the response will sound?

A. Make changes in the skill code, deploy it, and test it on the device.
B. Modify the SSML in the skill, save it, and check Amazon CloudWatch for errors.
C. Edit the SSML in the skill and use the Alexa Simulator tab on the Test page of the developer console.
D. Use the Voice & Tone tab on the Test page in the developer console.

Answer 9

D

The most efficient way to test changes to the speed of a voice response is to use the Voice & Tone tab in the Alexa developer console. This allows for immediate testing of SSML modifications without requiring a full deployment and testing cycle on a device (A). Option B is incorrect because CloudWatch is for monitoring errors, not for real-time testing of voice response speed. Option C is less efficient because it only lets you test the modified SSML in the context of the entire skill response, whereas Option D focuses specifically on the voice and tone characteristics.

Question 10

An Alexa Skill Builder needs to change the invocation name of a new skill. What status should the skill be in to make this change?
A. In Development
B. Build
C. In Certification
D. Edit

Answer 10

A. In Development

The correct answer is A because the provided documentation and discussion indicate that invocation names are changed during the development phase of a skill. Option B (“Build”) is too late in the process; the invocation name would already be set. Option C (“In Certification”) is also too late, as changes are generally not allowed during certification. Option D (“Edit”) is not a recognized skill status in the Alexa Skill development lifecycle. The discussion comments support the “In Development” status as the appropriate stage for making this change.

Question 11

An Amazon Alexa interactive story skill needs to provide users with the option to resume the skill from where they left off when they last used the skill. Where should the data be stored to ensure that the skill will start at the correct location?
A. In a JSON file along with the skill’s AWS Lambda function
B. In Amazon DynamoDB
C. In the skill’s session object
D. In the skill’s request object

Answer 11

C

The correct answer is C because the skill’s session object is designed to store data specific to a user’s current interaction with the skill. This is a temporary store, however, and will persist only for the duration of that particular session. While DynamoDB (B) is a persistent data store and could be used, it is overkill for this purpose, and adds unnecessary complexity. Storing data within the Lambda function (A) is not appropriate; Lambda functions are stateless. The request object (D) contains information about the current user request and is not intended for persistent storage of session-specific data. The session object is the appropriate location to store the user’s progress for resuming their interactive story.

Question 12

An Alexa Skill Builder is developing a custom skill and needs to verify that the correct slot values are being passed into the AWS Lambda function. According to best practices, what is the MOST efficient way to capture this information?

A. Add a logging statement to write the event request to Amazon CloudWatch Logs.
B. Add an API call to write the environment variables to an Amazon S3 bucket when the function is invoked.
C. Add an API call to read the event information from AWS Cloud Trail logs and add a PutObject API call to write to an Amazon S3 bucket.
D. Add a statement to parse the JSON request and save to the local disk for the Lambda function

Answer 12

A

CloudWatch Logs provide a straightforward and efficient method for capturing and reviewing the event requests, including slot values, sent to the Lambda function. This avoids unnecessary API calls and data transfers.

Option B is incorrect because environment variables don’t contain the slot values; they hold configuration settings for the Lambda function. Option C is incorrect because CloudTrail logs system-level events, not the detailed request data needed for slot value verification. It’s also inefficient to involve S3. Option D is incorrect because Lambda functions are ephemeral; data written to the local disk is lost when the function execution ends. This is not a persistent solution.

Question 13

An Alexa Skill Builder submitted a child-directed skill for certification that lists the nearest skateboard parks. The Builder ensured that
* The skill does not link to an external account
* The skill stored the child’s preference by userId
* The skill has a valid privacy policy link in the skill Distribution page
Given this information, why will the skill fail certification?
A. The Builder did not upload a privacy policy document
B. Child-directed skills cannot use location information
C. The child’s userId cannot be used because of Personally Identifiable Information (PII) restrictions.
D. The skill should not be child-directed as it can be used by children over 13 years old.

Answer 13

B

The correct answer is B because child-directed skills are prohibited from collecting or using location information. While options A and C are potential issues in skill development, they are not the reason given the information provided in the question. Option D is incorrect; the issue is not the age range of potential users, but rather the use of location data.

Question 14

According to Amazon Alexa best practices, how should an Alexa Skill Builder prevent unintentional requests against a skill’s backend when using AWS Lambda?
A. Ensure that the session ID provided by the request to Lambda is not already in use.
B. Rotate the Lambda ARN regularly to prevent others from using the service.
C. Retrieve the Application ID property from the request JSON and validate it against the Lambda environment variables.
D. Provide the Lambda trigger with the Application ID so that it validates on the ask trigger.

Answer 14

C

The correct answer is C because validating the Application ID against the Lambda environment variables is the recommended method to ensure that requests originate from your Alexa skill. Option A is incorrect because session IDs are not designed for this type of security validation. Option B is incorrect because rotating the Lambda ARN doesn’t prevent unauthorized requests; it only changes the access point. Option D is incorrect because while the Application ID is crucial, providing it to the trigger itself doesn’t inherently provide validation; the skill’s backend logic must actively validate it, which is what option C describes.

Question 15

An Alexa Skill Builder would like to improve a skill’s help experience by leveraging the user’s activity leading up to the help request to contextualize the help response. Where should the skill obtain the necessary context?
A. Load the user’s recent activity from the Intent Request History API, then use this to provide context to the AMAZON.HelpIntent request.
B. Retrieve the recent activity from the context object passed with the AMAZON.HelpIntent request.
C. Use a session attribute to store the intent name for each request, then use this to provide context to the AMAZON.HelpIntent request.
D. Retrieve the recent activity from the slot values passed with the AMAZON.HelpIntent request.

Answer 15

C

The correct answer is C because session attributes provide a mechanism to store and retrieve information across multiple requests within a user’s session. Storing the intent name in a session attribute allows the skill to recall the user’s previous actions and tailor the help response accordingly. This approach is explicitly supported by Amazon’s documentation and community examples, as shown in the provided discussion.

Option A is incorrect because the Intent Request History API is not designed for real-time contextual help within a single session. It provides historical data, not immediate context. Option B is incorrect because the context object associated with AMAZON.HelpIntent doesn’t inherently contain the user’s recent activity history. Option D is incorrect because slot values only reflect the data provided within a single intent request and do not provide a history of previous user interactions.

Question 16

An Alexa Skill Builder adds a colleague to a skill using the beta test feature. The colleague logs in to the developer console to edit the interaction model and cannot see the skill. Why is this happening?
A. The colleague needs the ROLE_ADMINISTRATOR enablement.
B. The skill was not submitted for publishing.
C. The colleague was not made an administrator in the beta test tool.
D. The colleague has not been added to the skill’s developer account.

Answer 16

C. The colleague was not made an administrator in the beta test tool.

The discussion highlights that simply adding a colleague’s email address isn’t sufficient; the colleague must be specifically granted administrator privileges within the beta testing tool itself. Option A is incorrect because ROLE_ADMINISTRATOR is not relevant to beta testing. Option B is irrelevant; beta testing happens before publishing. Option D is also incorrect because the process described in the linked blog post implies that the colleague’s developer account is added, but they lack the necessary beta testing permissions.

Question 17

What are the prerequisites for implementing account linking for Amazon Alexa smart home skills?

A. OAuth 2.0 with either implicit grant flow or authorization code grant flow
B. OAuth 2.0 with authorization code grant flow
C. OpenID Connect with JSON.Web Token (JWT)
D. OAuth 1.0/2.0 with implicit grant flow

Answer 17

B

The correct answer is B because Amazon’s documentation specifically states that for smart home skills, the authorization code grant type within OAuth 2.0 is required for account linking. Option A is incorrect because while implicit grant flow is acceptable for some types of Alexa skills, it’s not required and not sufficient for smart home skills. Option C is incorrect because it specifies OpenID Connect and JWT, which are not the required protocols for this use case. Option D is incorrect because it includes OAuth 1.0, which is not supported; only OAuth 2.0 is.

Question 18

An Amazon Alexa trip planner skill includes a PlanMyTripIntent with slots for fromCity, toCity, departDate, and returnDate. A user initiates a trip plan (“Ask plan my trip to start a new trip leaving from Seattle.”), and Alexa confirms the departure city. The user confirms, and Alexa saves the trip, then asks: “I’ve saved your trip. Do you want to create another trip?” The user replies “Yes”. What intent is invoked by the user’s final “Yes”?

A. PlanMyTripIntent
B. AMAZON.CancelIntent
C. AMAZON.YesIntent
D. LaunchRequest

Answer 18

C. AMAZON.YesIntent

The user’s final “Yes” is a response to a general question outside the context of the PlanMyTripIntent. Since the trip planning is complete, this “Yes” doesn’t relate to the specifics of the trip planning intent. Therefore, it invokes the generic AMAZON.YesIntent. Option A is incorrect because the trip planning is already finished. Option B is incorrect because the user isn’t canceling anything. Option D is incorrect because LaunchRequest is invoked when the skill is initially launched, not in response to a question during a user interaction.

Question 19

An Alexa Skill Builder implemented in-skill purchasing for premium content using an entitlement product deployed via the ASK CLI. Testing the BuyInskillProductIntent in the developer console yields the error: “Sorry, this product is not available with your current language setting.” How can this error be resolved?

A. Change the release date and redeploy the product.
B. Change the locale in the Alexa Simulator tab in the developer console.
C. Change the AWS Lambda function to include the correct locale in the Connections.SendRequest directive.
D. Change the language in the premium content.

Answer 19

C

The error message indicates a mismatch between the skill’s language setting and the product’s available languages. While changing the locale in the simulator (B) might seem like a solution, the problem lies in how the skill requests the purchase. The Connections.SendRequest directive within the AWS Lambda function handles the communication with the in-skill purchasing system. Therefore, ensuring the correct locale is included in this directive (C) is the correct way to resolve the issue. Option A is incorrect because the release date doesn’t affect language support. Option D is incorrect because changing the content language doesn’t address the in-skill purchasing system’s language requirements. The in-skill purchasing system needs to know which language the user is using to provide the correct product.

Question 20

An Alexa Skill Builder is building an order reporting skill. Occasionally, users need to enter 30-digit serial codes. How can this be accomplished while providing a good voice user interface experience?
A. Manually extend the timeout so that users can input all the numbers.
B. Use multiple requests for smaller segments of the code and store the data in session attributes.
C. Enter single digits one request at a time.
D. Request that users send the number using the Amazon Alexa app

Answer 20

B. Use multiple requests for smaller segments of the code and store the data in session attributes.

This is the best option because it breaks down a complex task (entering a 30-digit code) into smaller, more manageable chunks. This improves the user experience by avoiding lengthy input sessions and potential errors. Storing the data in session attributes allows the skill to remember the partial code across multiple turns of the conversation.

Option A is poor because manually extending timeouts doesn’t address the core issue of the difficulty of verbally inputting such a long code accurately. Option C is also inefficient and prone to errors. Option D is a poor user experience as it forces users away from the voice interface.

Question 21

An Alexa Skill Builder is developing a skill containing a multi-turn dialog that can be invoked with or without a specific intent request. On invocation, the skill needs to retrieve persistent attributes that have been saved by a previous invocation, and then copy them into session attributes. How should the Builder implement this functionality?

A. Place logic within the skill’s SessionEndedRequest intent handler to copy the persistent attributes into the session attributes.
B. Implement ResponseInterceptor containing logic that takes the current persistent attributes and copies them into session attributes.
C. Include logic within the LaunchRequest intent handler to retrieve persistent attributes and copy them into session attributes.
D. Implement RequestInterceptor containing logic which for new sessions, retrieves persistent attributes and copies them into session attributes.

Answer 21

D

The correct answer is D because RequestInterceptors are triggered at the beginning of each request, allowing the retrieval and copying of persistent attributes into session attributes specifically for new sessions. Option A is incorrect because SessionEndedRequest handles the end of a session, not the beginning. Option B is incorrect because ResponseInterceptors run after the response is generated, too late to affect the session attributes. Option C is incorrect because while LaunchRequest handles skill invocation, it doesn’t distinguish between new and existing sessions, potentially overwriting attributes unnecessarily for returning users. Therefore, only a RequestInterceptor can conditionally handle the attribute copying based on session status, making it the most appropriate choice.

Question 22

An Alexa Skill Builder has built a new custom skill backed by an AWS Lambda function. The Lambda function executes successfully from the Lambda console, however, the Lambda function cannot be successfully invoked in the developer console or from an Amazon Alexa enabled device. No error messages show in the function’s Amazon CloudWatch Logs. The Builder confirmed the endpoint has the correct ARN. What is likely causing this issue and how can it be corrected?

A. The Lambda application code has a bug that is causing it to crash. Modify the code to fix the bug, then redeploy the Lambda function.
B. The ASK SDK was not deployed with the Lambda function. Add the ASK SDK, then redeploy the Lambda function.
C. The ASK trigger for the Lambda function has been restricted to the wrong skill ID. Re-create the trigger with the correct skill ID.
D. The Lambda role does not have the correct AWS IAM permission. Update the IAM role associated with the Lambda function.

Answer 22

C

The most likely cause is that the ASK trigger for the Lambda function is restricted to the wrong skill ID (option C). Since the Lambda function works from the console but not from Alexa or the developer console, the problem is likely with the invocation process, not the function’s code itself. The ASK trigger verifies the skill ID before invoking the Lambda function; an incorrect skill ID prevents invocation without generating errors in CloudWatch logs. Options A, B, and D are less likely because the function works correctly when invoked directly from the Lambda console. Option A would result in errors in CloudWatch logs. Option B is unlikely given that the function is successfully called from the Lambda console, where SDK presence wouldn’t matter. Option D would likely prevent the Lambda function from working at all.

Question 23

An Alexa Skill Builder is developing a custom skill to play a live audio stream. What two built-in intents are required to implement the AudioPlayer interface?

A. AMAZON.NextIntent and AMAZON.StopIntent
B. AMAZON.ResumeIntent and AMAZON.PauseIntent
C. AMAZON.CancelIntent and AMAZON.ResumeIntent
D. AMAZON.RepeatIntent and AMAZON.StopIntent

Answer 23

B. AMAZON.ResumeIntent and AMAZON.PauseIntent

The correct answer is B because the Amazon developer documentation and community discussion confirm that AMAZON.ResumeIntent and AMAZON.PauseIntent are the two built-in intents required for implementing the AudioPlayer interface. These intents allow the user to pause and resume the audio stream. Options A, C, and D include intents that are not directly required for basic audio playback control via the AudioPlayer interface, although they might be used in a more complex skill.

Question 24

An Alexa Skill Builder is using session attributes to maintain a user’s state. What can the Builder do to ensure that a user’s session is not lost if they take too long to answer a question and the skill exits?
A. Set shouldEndSession to false in the response object to prevent the skill from exiting.
B. Handle the SessionEndedRequest request type and persist the user’s session to a database.
C. Return false from the SessionEndedRequest handler so the session does not exist.
D. Return a reprompt in the response object from the SessionEndedRequest handler.

Answer 24

B

The correct answer is B because the question describes a scenario where the session has already ended (due to the user taking too long to respond). Option A is incorrect because setting shouldEndSession to false only prevents the session from ending during a response; it doesn’t recover a session that has already ended. Option C is incorrect because returning false from SessionEndedRequest would likely not prevent data loss and may not even be a valid action in this context. Option D, while good practice for prompting users for further input during an active session, doesn’t address the data preservation issue after the session has already ended. Only persisting the session data to a database (option B) ensures that the information isn’t lost when the session times out.

Question 25

On Amazon Alexa enabled devices with a screen, which of the following standard built-in intents are handled by Alexa on the skill's behalf and are not forwarded to the skill? (Choose three.) A. AMAZON.NextIntent B. AMAZON.MoreIntent C. AMAZON.ScrollDownIntent D. AMAZON.PreviousIntent E. AMAZON.StartOverIntent F. AMAZON.NavigateHomeIntent

Answer 25

B, C, F

Question 26

An Alexa Skill Builder needs to set up an Amazon Alexa skill beta test. What user identifier should be used to add beta testers? A. AWS account number B. Alexa user email address C. Amazon vendor ID D. Amazon customer ID

Answer 26

B The correct answer is B, Alexa user email address. The discussion section unanimously indicates that the email address associated with the Alexa account is the correct identifier for adding beta testers. Options A, C, and D are incorrect because they are not user identifiers directly tied to an Alexa account and its usage. AWS account numbers identify AWS accounts, not specific Alexa users. Vendor and customer IDs are related to Amazon business accounts and are not relevant to identifying individual Alexa users for beta testing.

Question 27

An Alexa Skill Builder wants a skill to inform users of the number of times they have launched the skill. What approach should the Builder take to track this information? A. Store and increment a `launchCount` session attribute. B. Using the ASK SDK, implement a persistent attribute that is stored and retrieved from Amazon DynamoDB. C. Perform a scan of the Amazon DynamoDB table for the number of records and count each record as a skill launch. D. Set up a global variable in the code project that increments each time the code is invoked by Amazon Alexa.

Answer 27

B The correct answer is B because using a persistent attribute stored in DynamoDB is the most appropriate method for tracking the number of times a user has launched a skill. Session attributes (A) are temporary and will be lost when the session ends. Scanning the entire DynamoDB table (C) is inefficient and inappropriate for this purpose; it's designed for querying, not tracking individual user launches. A global variable (D) is also inappropriate because it would track all launches across all users, not individual user counts. Persistent attributes, as described in the provided reference link, offer a suitable solution for storing and retrieving user-specific data persistently across sessions.

Question 28

An Alexa Skill Builder is trying to enable a skill in the developer console. The skill model builds without errors, but the Builder is unable to turn on testing. What could be causing this? A. The skill's slot types are invalid. B. The skill does not have an example phrase. C. The skill does not have an invocation name. D. The skill does not have an endpoint.

Answer 28

C The correct answer is C because the documentation states that an invocation name (skill call name) is required to enable testing. Option A is incorrect because while invalid slot types might cause other issues, they don't prevent enabling testing. Option B is incorrect because while example phrases (sample utterances) are important for testing and the skill's functionality, they are not a prerequisite for simply turning on testing in the developer console. Option D is incorrect because an endpoint is required for the skill to function, but the ability to turn on testing is a separate issue that occurs *before* the endpoint is used; the developer console should allow enabling testing even before the skill is fully deployed.

Question 29

An Alexa Skill Builder published a skill and now wants to update it. Which of the following changes will make it necessary to have the skill re-certified? (Choose two.) A. Adding a new intent in the interaction model B. Changing the text of a skill response C. Adding a sentence to the skill description D. Changing the URL of the image on a skill card E. Updating the backend to point to a new database

Answer 29

A, D A is correct because adding a new intent changes the interaction model, requiring re-certification. D is correct because changing the skill's icon (image on a skill card) requires re-submission for certification. B is incorrect because changing the text of a skill response is a backend change and does not require re-certification. C is incorrect because adding to the skill description, while altering the skill's information, does require re-certification according to the provided resources. E is incorrect because updating the backend does not require re-certification.

Question 30

An Alexa Skill Builder is designing a skill with an intent that requires six slots. It's unlikely a user will provide all slot values in a single utterance. To handle this multi-turn conversation and have Alexa elicit any missing slots *without* explicitly specifying each slot in the backend code, what should the Builder do in the developer console? A. Keep track of filled slots in session attributes, and in the backend code, prompt the user for missing slots using `Dialog.ElicitSlot`. B. Mark the six slots as required, fill in the necessary prompts, and in the backend code, use the `Dialog.Delegate` directive until all slots are filled. C. Mark the six slots as required, fill in the necessary prompts, and in the backend code, use the `Dialog.ConfirmSlot` directive until all slots are filled. D. Mark the six slots as required, fill in the necessary prompts, and in the backend code, use the `Dialog.ElicitSlot` directive until all slots are filled.

Answer 30

B The correct answer is B because using `Dialog.Delegate` allows the Alexa system to manage the multi-turn dialog. The developer defines prompts for each slot in the interaction model. `Dialog.Delegate` hands off control to Alexa to determine which slot to elicit next based on the user's input and the defined prompts. Options A, C, and D require explicit handling of each slot in the backend code, which contradicts the requirement of *not* specifying each slot in the backend. Option C is incorrect because `Dialog.ConfirmSlot` is used to confirm a specific slot value, not elicit missing values. Option A is also incorrect because it requires the developer to manage the elicitation process, which is what the question aims to avoid. Option D is similar to A, handling elicitation manually instead of delegating to Alexa.

Question 31

An Alexa Skill Builder noticed that a large percentage of a food ordering skill's customers are not completing their transactions. The Builder needs to know what portion of the customers are leaving the skill by not responding, compared to the portion of customers who receive an error. Which report inside the Analytics section of the developer console will provide this information? A. Session Type Distribution B. Unique Customers Per Intent C. Average Session Per Customer D. Failed Utterances Per Intent

Answer 31

A. Session Type Distribution The correct answer is A because the Session Type Distribution report shows the different ways a session can end, including timeouts (customer not responding) and errors. Options B, C, and D are incorrect. "Unique Customers Per Intent" shows customer usage of specific intents, not session completion rates. "Average Session Per Customer" provides an average session length, not the reason for session termination. "Failed Utterances Per Intent" focuses on errors in the user's requests, not overall session failures due to lack of response or other errors.

Question 32

An Alexa Skill Builder is developing a complex skill to help users purchase items. The Builder has implemented the default AMAZON.HelpIntent, but when users ask for help they always reach the same unhelpful result. How should the Builder enhance the experience to guide the users toward the end goal? A. Check the state of the user conversation when AMAZON.HelpIntent is received to give contextual responses. B. Implement AMAZON.FallbackIntent to give a response when an utterance is not understood. C. Extend the default AMAZON.HelpIntent with more sample utterances. D. Implement a custom intent users can request when they need help.

Answer 32

A The correct answer is A because checking the conversation state allows the skill to provide help tailored to the user's current position in the purchasing process. Option B is incorrect because AMAZON.FallbackIntent handles situations where the skill doesn't understand the user's input, not requests for help. Option C is incorrect because adding more sample utterances to the default AMAZON.HelpIntent doesn't address the core problem of providing context-specific assistance. Option D is incorrect because while a custom intent *could* work, it requires users to remember a specific command for help, which is less intuitive than having contextually relevant help readily available through the standard AMAZON.HelpIntent.

Question 33

AMAZON.MoreIntent is comparable to which of the following Amazon Alexa intents? (Choose two.) A. AMAZON.ResumeIntent B. AMAZON.ScrollDownIntent C. AMAZON.RepeatIntent D. AMAZON.ScrollRightIntent E. AMAZON.LoopOnIntent

Answer 33

B and D The correct answers are B (AMAZON.ScrollDownIntent) and D (AMAZON.ScrollRightIntent). The AMAZON.MoreIntent is designed to display more content, mirroring the functionality of scrolling down or to the right. Option A (AMAZON.ResumeIntent) is incorrect because it relates to resuming a paused activity, not displaying more content. Option C (AMAZON.RepeatIntent) is incorrect because it deals with repeating previous output, not accessing more content. Option E (AMAZON.LoopOnIntent) is unrelated to displaying more content.

Question 34

An Alexa Skill Builder did not include a display template in a skill. When the skill is used with an Amazon Alexa enabled device with a screen, cards used in the skill are: A. rendered as a gray screen. B. rendered using the skill icon as the foreground image. C. rendered using the skill icon as the background image. D. rendered using BodyTemplate1.

Answer 34

D The correct answer is D because if a display template is not specified, Alexa defaults to using BodyTemplate1 to render cards on devices with screens. Options A, B, and C are incorrect because Alexa does not render a blank gray screen, nor does it automatically use the skill icon as the foreground or background image in the absence of a specified display template. The default behavior is to use the standard BodyTemplate1.

Question 35

An Alexa Skill Builder is developing a skill that must send an initial welcome email to each new user. What combination of features would the Builder use to satisfy this requirement? (Choose two.) A. Alexa Settings API B. Entity resolution C. Customer Profile API D. Device Address API E. Data persistence

Answer 35

C and E The correct answer is C and E because: * **C. Customer Profile API:** This API allows access to user information, including their email address, which is necessary to send a welcome email. * **E. Data Persistence:** This is needed to track whether a welcome email has already been sent to a specific user. Without data persistence, the skill would send a welcome email every time the user interacts with it. The other options are incorrect because: * **A. Alexa Settings API:** This API deals with user settings within Alexa, not user contact information. * **B. Entity Resolution:** This is used for identifying and linking related entities within a dataset; it is not relevant for sending emails. * **D. Device Address API:** This API provides the device's address, not the user's contact information.

Question 36

An Alexa Skill Builder built a skill using AWS Lambda. The Lambda function works when running the code on a local machine with a runtime of 4.5 seconds, but during skill testing, the Builder receives an error response. Which collection of steps will address the issue? (Choose two.) A. Change the Amazon Alexa default timeout to 5 seconds. B. Change the default timeout of the Lambda function to 5 seconds. C. Call the Progressive Response API and send a directive to reduce latency. D. Increase the size of the memory allocated to the Lambda function. E. Clone the Lambda function to another AWS Region.

Answer 36

B and D The correct answers are B and D because: * **B. Change the default timeout of the Lambda function to 5 seconds:** The Lambda function is exceeding its timeout limit. Increasing the timeout to a value greater than 4.5 seconds will allow the function to complete successfully. * **D. Increase the size of the memory allocated to the Lambda function:** Increasing the memory allocated to the Lambda function can improve its performance and reduce execution time. This could resolve the timeout error if the function's performance bottleneck is related to memory limitations. Option A is incorrect because the Alexa skill's timeout is irrelevant; the Lambda function's timeout is the primary issue. Option C is incorrect because the Progressive Response API is used for handling long-running requests and providing interim responses to the user; it doesn't directly address the timeout error. Option E is incorrect because cloning the function to another region won't inherently solve the performance or timeout issue. The problem likely resides within the function's code or resource allocation, not its location.

Question 37

Which of the following are ways that a user can interact with an Amazon Alexa skill using both voice and non-voice methods? (Choose three.) A. Scrolling through a list B. Selecting an action C. Zooming in on an image D. Pausing a video E. Muting the device F. Pairing with Bluetooth

Answer 37

D, E, F Pausing a video, muting the device, and pairing with Bluetooth can all be done both vocally (through voice commands to Alexa) and non-vocally (through a physical button or app interface). Options A, B, and C are primarily non-voice interactions; while some Alexa skills *might* offer voice control over these actions, it isn't a consistent or standard feature across all skills.

Question 38

An Alexa Builder is working on a skill for music streaming. When a user says, "Alexa, stop" the skill needs to know where the user was in the song, so that when returning to the skill in a new session, the song can pick up where it left off. Which section of the following JSON [Image](https://www.examtopics.com/assets/media/exam-media/03310/0003400001.png) contains the data indicating where the song left off? A. session.attributes B. Context.AudioPlayer C. session.user D. context.System.device.supportedInterfaces.AudioPlayer

Answer 38

B The correct answer is B, Context.AudioPlayer. The discussion clarifies that the offsetInMilliseconds parameter within the Context.AudioPlayer payload stores the information about the song's playback position. Option A (session.attributes) is incorrect because while session attributes can store data, they are not specifically designed for tracking audio playback. Option C (session.user) stores user-specific information, not playback state. Option D (context.System.device.supportedInterfaces.AudioPlayer) describes the device's capabilities, not the current playback state.

Question 39

The namespace value in the header of the incoming directive for an Amazon Alexa smart home skill specifies the: A. context of the message B. capability interface of the message C. endpoint specified in the message D. control message for the directive

Answer 39

B. capability interface of the message The discussion clarifies that the correct answer is B, citing the Amazon developer documentation. The namespace indicates which capability interface the directive relates to (e.g., "Alexa.PowerController"). Option A is incorrect because the context is conveyed through other parts of the message, not solely the namespace. Option C is incorrect as the endpoint is specified elsewhere in the request. Option D is incorrect because the namespace does not define the type of control message; that's determined by the `name` field within the directive.

Question 40

An Amazon Alexa skill fetches data from a third-party API. The API response time is variable, sometimes taking up to 5 seconds. What is the recommended method to inform users that the skill is processing their request and hasn't failed? A. Prefetch the data expected by the skill from the third-party API using Amazon CloudWatch Events. B. Call the Progressive Response API and send a directive, such as VoicePlayer.Speak. C. Ask a follow-up question for clarification to engage the user while waiting for the initially requested response. D. Respond to the user stating that the data will be ready soon, and upon the next launch of the skill, provide the user with the response they initially requested.

Answer 40

B The correct answer is B because using the Progressive Response API allows the skill to send interim responses to the user while it's waiting for the API call to complete. This provides immediate feedback, assuring the user that the skill is working. Option A is incorrect because prefetching data using CloudWatch Events doesn't address the real-time user experience; it's more suitable for background tasks or scheduled updates. Option C is incorrect because asking follow-up questions can be disruptive and doesn't directly address the delay. Option D is incorrect because delaying the response until the next skill launch is a poor user experience.

Question 41

A travel booking skill has slot elicitation for `fromCity`, `toCity`, and `travelDate` inside of a dialog. After going live, the skill is getting negative reviews stating that the skill does not understand the city names when customers try to book travel. The Alexa Skill Builder adds `AMAZON.FallbackIntent` to the interaction model with the goal of providing better messaging for out-of-domain utterances. Why will the addition of `AMAZON.FallbackIntent` fail to resolve the customer issue? A. AMAZON.FallbackIntent will not be triggered in the middle of a dialog. B. A handler cannot be added to a Live skill. C. AMAZON.FallbackIntent will cause the slot values to be incorrectly mapped. D. AMAZON.FallbackIntent cannot be present in a skill with a Dialog directive.

Answer 41

A

Question 42

An Alexa Skill Builder is using Amazon S3 to stream large quantities of static audio and video content throughout the world with an Amazon Alexa skill. Which additional AWS service will help the Builder decrease latency and improve the reliability of the streaming media content for the global audience? A. AWS Cloud9 B. Amazon Kinesis C. Amazon Route 53 D. Amazon CloudFront

Answer 42

D. Amazon CloudFront Amazon CloudFront is a Content Delivery Network (CDN) that improves latency and reliability of content delivery by caching content closer to users geographically. Options A, B, and C are incorrect. AWS Cloud9 is an integrated development environment (IDE). Amazon Kinesis is a service for real-time data streaming, not suited for static content delivery. Amazon Route 53 is a DNS service for routing traffic, but doesn't directly improve content delivery speed or reliability like a CDN.

Question 43

An Alexa Skill Builder needs to have knowledge of the previous prompt that was presented to the user in order to give context to the user's response. How can the Builder accomplish this? A. Find the corresponding prompt using the list events feature B. Store a reference to the prompt used as a session attribute C. Call the Intent Request History API to identify which prompt was used. D. Parse the context object from the skill request.

Answer 43

B. Store a reference to the prompt used as a session attribute. The most likely correct answer is B because session attributes are designed to persist data across multiple turns of a conversation with an Alexa skill. Storing the previous prompt as a session attribute allows the skill to easily access and utilize this information to provide context to the user's response. Option A is incorrect because the list events feature is not designed for accessing the previous prompt in a conversation; it's intended for different purposes like tracking skill usage. Option C is incorrect because the Intent Request History API is not directly used for this purpose; it's generally used for debugging and analysis. Option D is incorrect; while the context object may contain some relevant information, it's not the primary or most efficient method for managing the previous prompt within the skill's internal state. The session attribute is the designated and reliable method for this task.

Question 44

An Alexa Skill Builder is developing a skill that enables users to purchase train tickets. The Builder wants to give users the ability to modify the departure time if they are misunderstood, before the skill proceeds to purchase the tickets. Which option should be used to implement this functionality within the session? A. Implement AMAZON.FallbackIntent so the user can stop the execution when they have been misunderstood. B. Implement AMAZON.StopIntent so the user can stop the execution and implement reprompt with a new invocation of the skill. C. Implement AMAZON.CancelIntent so the user can cancel the order, then set shouldEndSession to false, and prompt the user for the next action. D. Implement the welcome message, providing clear instructions to the user describing how to format an order.

Answer 44

C The correct answer is C because AMAZON.CancelIntent allows the user to cancel the current order (the misunderstood departure time) without ending the session. Setting `shouldEndSession` to false keeps the skill active, allowing the user to provide the correct information and continue with the purchase. Option A is incorrect because AMAZON.FallbackIntent is for handling unexpected user input, not specifically for correcting a part of the order. Option B is incorrect because AMAZON.StopIntent ends the session entirely, requiring the user to restart the skill. Option D is a preventative measure that might reduce misunderstandings but does not directly address correcting a misunderstood order within the current session.

Question 45

An Alexa Skill Builder is building an interactive storytelling skill where the user can choose their own path through the story. The Builder wants Amazon Alexa to narrate the story and use custom audio for the dialogue for each character, along with various sound effects. How can these requirements be met? A. Record custom audio for each segment of the story, then stream audio files stored on Amazon S3 using the AudioPlayer interface for each segment of the story. B. Develop each response as text, and let Alexa speak each part differently using SSML markup for to alter the Alexa voice for each character. C. Record custom audio for each character voice and the sound effects in the story, then include the custom audio using SSML markup in the response. D. Record the entire story as custom audio and stream the audio stored on Amazon S3 using the AudioPlayer interface.

Answer 45

C The correct answer is C because it allows for the use of custom audio for character voices and sound effects, which are crucial for creating an immersive interactive storytelling experience. SSML (Speech Synthesis Markup Language) provides the mechanism to seamlessly integrate these pre-recorded audio files into the Alexa response. Option A is incorrect because while it uses custom audio, it streams entire segments. This approach lacks the flexibility to dynamically adjust the story based on user choices, which is a core requirement of an interactive story. Option B is incorrect because using only `` within SSML offers limited control over character voices. It primarily manages effects like whispering, not full voice customization. Option D is incorrect because it records the entire story upfront, negating the interactive nature of the skill. Users wouldn't be able to make choices that affect the narrative flow.

Question 46

An Alexa Skill Builder is interested in gathering information about returning users. What analytics tool in the developer console would provide the Builder with this information? A. Customers B. Interaction Path C. Sessions D. Retention

Answer 46

D. Retention The correct answer is D because the provided discussion explicitly states that the "cohort chart in the Retention section of the Analytics page shows the number of returning users". Option A, Customers, is incorrect because it provides general customer information, not specifically about returning users. Option B, Interaction Path, shows the flow of user interactions, not specifically returning users. Option C, Sessions, shows the number of user sessions, but doesn't differentiate between new and returning users.

Question 47

During testing of a new Amazon Alexa skill, the skill repeatedly fails and invokes the function defined in the `addErrorHandler` method specified on the SkillBuilder object. Amazon CloudWatch Logs indicate that the failure occurs whenever `AMAZON.HelpIntent` is received. How should this error be corrected? A. `AMAZON.HelpIntent` should be handled by the SDK. The Builder should raise a support ticket with Amazon. B. The Builder should ensure that the intent handler is coded so that it tests for `AMAZON.HelpIntent` in its `canHandle` method, and when detected, returns `true`. C. The Builder should add logic to provide help instructions to the function defined in the `addErrorHandler` method specified on the SkillBuilder object. D. The Builder should add an `AMAZON.HelpIntent` entry to the interaction model to ensure the request for help is recognized by the skill.

Answer 47

B The correct answer is B because the `addErrorHandler` function is being invoked, indicating that the skill is not handling the `AMAZON.HelpIntent`. Option B correctly addresses this by suggesting that the developer ensure their intent handler explicitly checks for and handles this built-in intent using the `canHandle` method. Option A is incorrect because the SDK doesn't automatically handle all intents; the developer needs to explicitly handle built-in intents like `AMAZON.HelpIntent`. Option C is incorrect because the error handler shouldn't contain the help logic; the intent handler should. Option D is incorrect because the interaction model likely already includes `AMAZON.HelpIntent`; the problem is the skill's failure to properly handle it within its code.

Question 48

An Alexa Skill Builder has created a taxi hiring skill. The skill needs to find out when the customer wants a taxi, where the customer is traveling from, and where the customer wants to go. The Builder is currently asking each question individually, in the following order: "Where do you want to take a taxi from", "Where do you want to take a taxi to", "When do you need a taxi". To ensure the voice interaction is flexible, how should this information be gathered regardless of the order in which the user provides it? A. Use a single intent and slot. Inspect the incoming slot value and categorize then response, then prompt for the remaining information. B. Create an intent for each question and include slots for each piece of information on every intent C. Create three intents with one slot each. Use Dialog.ElicitSlot to fill the slots. D. Create a single intent with three slots. Use the Dialog.Delegate directive to fill the slots.

Answer 48

D The correct answer is D because using a single intent with three slots allows the user to provide the information in any order. The `Dialog.Delegate` directive handles the conversation flow, ensuring all three slots are filled before the skill proceeds. Option A is incorrect because manually categorizing responses within a single slot is cumbersome and error-prone. It doesn't offer flexibility in the order of information provided. Option B is incorrect because creating separate intents for each piece of information restricts the user from providing the information in a flexible order; the user must follow a strict order. Option C is incorrect because while using `Dialog.ElicitSlot` helps manage the conversation flow, having three separate intents still doesn't allow the user to provide the information flexibly. The user is still required to answer questions in the order defined by the three separate intents.

Question 49

Which of the following occur when a beta test of a live skill times out? (Choose two.) A. The beta tester will lose access to the beta skill and will need to reenable the live skill B. The beta tester will receive an email saying the beta test has ended C. The administrator will receive an email confirming that the beta tester has been removed D. The beta tester will lose access to the beta skill but will maintain access to the live skill E. The beta skill history in the Amazon Alexa app will disappear once the beta test has ended

Answer 49

A and B The discussion section shows conflicting answers, however, the response from "originaly" referencing Amazon's official documentation (linked) is considered the most reliable. Option A is correct because the beta tester loses access to the beta skill and must re-enable it to use it again. Option B is correct because the tester receives an email notifying them of the beta test's end. Option C is incorrect because there's no mention of an email to the administrator. Option D is incorrect because the tester loses access to the beta skill. Option E is not explicitly stated in the provided text.

Question 50

When testing an Amazon Alexa skill using the Test page in the developer console, the JSON output is null. What is the MOST likely cause? A. The AWS Lambda function is encountering an error and is not returning a result. B. The endpoint ARN has not been configured. C. The ASK has not been granted access to trigger the AWS Lambda function. D. Skill ID verification has been disabled.

Answer 50

C. The ASK has not been granted access to trigger the AWS Lambda function. Although the discussion mentions option B, and some users suggest option C, the most likely cause of a null JSON response in the Alexa developer console's test page is that the ASK (Alexa Skills Kit) lacks the necessary permissions to invoke the AWS Lambda function. A configuration issue with the endpoint ARN (option B) would likely result in a different error message, rather than a simple null response. Option A (Lambda function error) would likely produce an error message within the Lambda logs or a different error response. Option D (Skill ID verification) is irrelevant to receiving a null JSON response. Therefore, a lack of necessary permissions (C) is the most plausible explanation for a null JSON output.

Question 51

An Alexa Skill Builder is using the AudioPlayer.Play directive and would like to automatically start the next audio track at the end of the current track. Which AudioPlayer playBehavior values will allow the Builder to achieve this without changing the audio currently playing? (Choose two.) A. REPLACE_ALL B. CLEAR_ENQUEUED C. REPLACE_ENQUEUED D. ENQUEUE

Answer 51

C and D The correct answers are C (REPLACE_ENQUEUED) and D (ENQUEUE). REPLACE_ENQUEUED replaces any queued audio with the new audio item, ensuring the next track plays after the current one finishes without interrupting the currently playing audio. ENQUEUE adds the next audio item to the queue, so it plays automatically after the current one completes. A (REPLACE_ALL) and E (CLEAR_ALL) are incorrect because they would replace or clear the currently playing audio, which the question specifies should not happen. B (CLEAR_ENQUEUED) is incorrect because it only clears any previously enqueued audio, but doesn't add the next track to be played.

Question 52

An Alexa Skill Builder implemented the built-in intent AMAZON.HelpIntent. In some cases, users are asking for help using phrases that are specific to a skill's terminology. Amazon Alexa does not understand these phrases are help requests, and they are not being routed to AMAZON.HelpIntent. According to best practices, how can this situation be corrected? A. Create custom intents using the help utterances that are specific to the skill, and remove AMAZON.HelpIntent. B. Use AMAZON.FallbackIntent to capture spoken phrases that do not match AMAZON.HelpIntent, then determine if the user needs help. C. Define custom slots for AMAZON.HelpIntent to capture the additional details in the users' help requests D. Extend the standard built-in AMAZON.HelpIntent using additional samples in the skill's interaction model

Answer 52

D The correct answer is D because extending the AMAZON.HelpIntent with additional sample utterances specific to the skill's terminology allows Alexa to recognize these phrases as help requests and route them appropriately. This leverages the existing built-in intent, avoiding the need for creating entirely new intents. Option A is incorrect because removing AMAZON.HelpIntent removes the standard help functionality, leaving users with no way to access help through standard phrases. Option B is incorrect because AMAZON.FallbackIntent is for handling utterances that don't match any defined intent, not specifically for expanding the functionality of AMAZON.HelpIntent. Option C is incorrect because custom slots are for capturing specific information within an intent, not for expanding the range of phrases that trigger an intent.

Question 53

An Alexa Skill Builder is developing a skill that tells users facts about their neighborhoods using their postal codes. When running the skill, users receive error messages from their Amazon Alexa devices. When the Builder checks the logs, error messages are being sent when the device location API is used to get users' postal codes. What are the potential causes of this issue? (Choose two.) A. The Alexa devices are registered in the incorrect postal code. B. The AWS Lambda function is not registered to the skill. C. The users have opted out of giving their location information. D. The application ID of the skill does not match the expected application ID of the AWS Lambda function. E. The skill is not registered to access information in the developer console.

Answer 53

C and E The correct answers are C and E because: * **C. The users have opted out of giving their location information:** If users have disabled location sharing, the skill will not be able to access their postal codes, resulting in errors. * **E. The skill is not registered to access information in the developer console:** The skill needs the necessary permissions granted in the developer console to access user location data through the device location API. Without these permissions, the API call will fail. Option A is incorrect because the skill uses the device's reported location, not a manually entered postal code. Option B is incorrect because a missing Lambda function registration would typically result in a different error, related to the Lambda function not being invoked at all, not specifically errors within the location API. Option D is incorrect because a mismatch in application IDs would prevent the skill from functioning entirely, not just the location-based functionality.

Question 54

An Alexa Skill Builder wants to create a skill that asks the user two yes/no questions: Alexa: Do you like cats? Alexa: Do you like dogs? When the user answers "yes", how should the Builder code the handler to know which question the answer refers to? A. Using session attributes, store the previous question as the context for use in the AMAZON.YesIntent handler. B. Within the AMAZON.YesIntent handler, prompt the user to repeat the name of the animal that they like. C. Within the AMAZON.YesIntent handler, define a slot to store and retrieve the previously asked question. D. Access Amazon CloudWatch Logs and retrieving the previous question topic from the recent log messages.

Answer 54

A The correct answer is A because using session attributes is the standard and most efficient way to track context within an Alexa skill. The AMAZON.YesIntent doesn't inherently know which question the "yes" refers to; the developer must explicitly manage this context. Session attributes provide a mechanism to store and retrieve this information within the skill's execution. Option B is inefficient and a poor user experience, requiring the user to repeat information. Option C is incorrect because you cannot define slots within the AMAZON.YesIntent. Option D is incorrect because accessing CloudWatch logs is not appropriate for managing runtime context within a skill; it's for debugging and monitoring.

Question 55

Before beginning the development of a skill, an Alexa Skill Builder wants to hear how a statement will be pronounced by Amazon Alexa in another locale. How should the Builder test the pronunciation of the statement? A. Use the Alexa Simulator tab in the developer console to interact with the skill using the selected locale. B. Create a developer account for the specific locale, and use a device to interact with the skill. C. Use the Voice & Tone tab in the developer console to play the statement using the selected locale. D. Use the Manual JSON tab in the developer console to create a sample JSON response

Answer 55

C The correct answer is C because the Voice & Tone tab in the Alexa developer console allows you to input text and hear how Alexa will pronounce it in various locales. Option A is incorrect because the simulator tests the skill's functionality, not just the pronunciation of a specific statement. Option B is incorrect as it is unnecessarily complex and resource-intensive; creating a new developer account and using a device is not needed for simply testing pronunciation. Option D is incorrect because the Manual JSON tab is for testing the skill's JSON responses, not for testing pronunciation.

Question 56

A skill contains a PlanMyTrip intent configured to require slots `fromCity` and `toCity`. It contains the following interaction: Alexa: What city are you leaving from? User: I'm leaving from Seattle. Alexa: You want to fly out of Seattle Tacoma International Airport, right? User: Yes - Which of the following should the Builder use to confirm the value of the `fromCity` slot only? (Choose two.) A. Use the `Dialog.ConfirmSlot` directive, including the full airport name in the `outputSpeech` object. B. Use the `Dialog.ConfirmIntent` directive, including the full airport name in the `updatedIntent` object. C. Use the `Dialog.Delegate` directive, including the full airport name in the `updatedIntent` object. D. Use the `Dialog.Delegate` directive, including the full airport name in the `outputSpeech` object. E. Use the `Dialog.ElicitSlot` directive, including the full airport name in the `updatedIntent` object.

Answer 56

A and D The correct answers are A and D. Option A is correct because `Dialog.ConfirmSlot` is specifically designed to confirm a single slot value. The `outputSpeech` object allows you to provide the text Alexa uses to confirm the slot's value, which includes the airport name. Option D is correct because `Dialog.Delegate` hands off control back to the dialog management system. While it doesn't directly confirm, including the airport name in `outputSpeech` allows for Alexa to present the information before proceeding with the next step in the dialog. Option B is incorrect because `Dialog.ConfirmIntent` confirms the *entire* intent, not a single slot. Option C is incorrect because `Dialog.Delegate` doesn't explicitly confirm anything; it simply delegates further processing. Option E is incorrect because `Dialog.ElicitSlot` is used to ask for a slot value, not confirm it.

Question 57

An Alexa Skill Builder is troubleshooting issues with a custom skill backed by an AWS Lambda function that integrates with an external API controlling a light bulb. The Builder observes that when saying "Alexa, turn on the light" the response is "light is not responding" and 10 seconds later, the light turns on. What is the MOST likely cause for this issue and how can it be solved? A. The Lambda function is not executing fast enough. Double the currently specified Lambda memory allocation in the Lambda basic settings section. B. The default Lambda function timeout setting is too short and the Lambda function times out before the response from the external API can be processed and a reply can be sent back to Amazon Alexa. Increase the Lambda timeout limit. C. There are too many concurrent Lambda functions running, causing the existing Lambda function to block and then time out before a response can be returned to Amazon Alexa. Increase the Lambda function reserve concurrency value to 30, then verify that the function can complete its work within 10 seconds. D. There is a bug in the Lambda function code preventing the external API from being called. Enable Lambda debugging and error handling and check Amazon CloudWatch Logs for the error, then modify the code accordingly.

Answer 57

B The correct answer is B because the problem description clearly states that the light turns on 10 seconds after the "light is not responding" message. This strongly suggests that the Lambda function is taking longer than its timeout setting to complete the request to the external API. Increasing the Lambda timeout will allow the function to complete its task and send a response to Alexa. Option A is incorrect because doubling the memory allocation might improve performance, but it doesn't address the core issue of the timeout. The delay is not due to slow execution, but the exceeding the time limit before execution is complete. Option C is incorrect because the problem description doesn't indicate high concurrency issues. The delay is consistent and not intermittent, which would be more indicative of concurrency problems. Option D is incorrect because while a bug in the Lambda function is possible, the consistent 10-second delay before the light turns on points towards a timeout issue rather than a code bug preventing the API call entirely. A code bug would likely result in a consistent failure to turn on the light, not a delayed response.

Question 58

An Alexa Skill Builder modified an AWS Lambda function serving as a skill's endpoint, resulting in an error upon skill launch. How can the Builder utilize the Lambda console to trigger the function and debug the code? A. Create a Lambda test event using the JSON request as input to find the specific error within the code. B. Create a Lambda test event using the JSON response as output to find the specific error within the code. C. Check the JSON response to see if there are any syntax errors in the code. D. Create a Lambda test event using the JSON interaction model to find the specific error within the code.

Answer 58

A The correct answer is A because the problem originates from the input to the Lambda function (the Alexa request). To debug, the builder needs to simulate this input using a test event with the JSON request data. The Lambda function will process this simulated request, and any errors occurring during processing will be revealed in the logs or execution results. Option B is incorrect because the response is the *output* of the Lambda function, not the source of the error. Examining the response is useful *after* identifying the problem with the input. Option C is incorrect because checking only the JSON response won't necessarily show errors within the Lambda function's code itself. Syntax errors in the response could be *caused by* problems in the function but won't pinpoint the root cause. Option D is incorrect because while the interaction model is part of the Alexa skill, it's not directly used to trigger or debug the Lambda function. The interaction model defines how the user interacts with the skill, but the Lambda function needs the raw JSON request data to process it.

Question 59

An Alexa Skill Builder has created a custom skill about basketball including a HowToPlayBasketball intent. When looking at the Intent History page in the developer console, the Builder sees that a number of users are asking the skill how to play baseball. The Builder wants to add a relevant response directing the user back to the topic of basketball. How should the Builder implement this? A. Add AMAZON.FallbackIntent and respond with a message about baseball in the handler B. Create a custom intent related to baseball, and when matched, provide a relevant response C. Add more sample utterances related to baseball in the HowToPlayBasketball intent D. Create a new custom baseball slot and add a slot-filling utterance to the HowToPlayBasketball intent.

Answer 59

A The correct answer is A because the AMAZON.FallbackIntent is designed to handle unexpected user requests. Since users are asking about baseball (unrelated to the skill's basketball focus), the FallbackIntent is the appropriate mechanism to catch these requests and provide a relevant response guiding them back to the intended topic of basketball. Option B is incorrect because creating a separate baseball intent would require significant additional development and wouldn't address the core issue of users mistakenly interacting with the basketball skill. Option C is incorrect because adding baseball utterances to the HowToPlayBasketball intent would make the skill respond to baseball questions as if they were about basketball, which is not desirable. Option D is incorrect because adding a baseball slot to the intent is irrelevant to the problem of handling unrelated user requests.