Harshit@Amazon avatar image
Harshit@Amazon posted

Tips for a Successful Alexa Skill Certification: Contextualized TTS or audio playback response

The speech output for Alexa is accomplished via text-to-speech (TTS) technology that converts sequences of words into natural sounding, intelligible audio responses. Hence, it is important that the skill's example phrases reach the skill. In order to create a good customer experience, you should make sure the example phrases are working correctly and are error free.

Note - All customer facing example phrases must function without error since these are the interactions that customers are most likely to try.

Issue 1) At times, some launch phrases work better than others depending on the invocation name.

You should always confirm your example phrase by voice and check the utterance history in the Alexa App. Here is how you can check the Utterances in the History page of the Alexa Web app:


The history page shows what utterance is registered in your request. Here you need to validate if the the example phrase that you have used to invoke the skill, match the phrase present in the skill model . To validate if the example phrases or launch phrases match the message you see in the following format: “This request was sent to <skill name>.”

If the skill does not launch correctly, you will see it in the above option if the request successfully went to your skill OR the skill returned a non- contextualized response or error and does not reflect the correct skill name.

Issue 2) A lot of times the skill response is different from the intended request or different than what’s described or written in the skill’s description and hence changes are that the skill will fail certification.

Here, you need to make sure of the following details:

The example phrases respond to the intended requests and matches to what is written in the skill description.

Good example:

User: Alexa, open radio XYZ

Skill: Welcome to radio XYZ /Here’s is radio XYZ .........<Plays audio>

Bad example:

User: Alexa, open bird sound

Skill: Welcome to bird sound! What would you like to do?

User: Play woodpecker sound

Skill: <Plays a dog sound>

In order to make good customer experience, having the example phrases intend to work as expected and as accurate to the skill’s character and description significantly increases customer interactions.

Issue 3) Make sure that all example phrases respond appropriately without errors or irrelevant responses. Often example phrases in certain cases return an error response even though it reaches the skill correctly. The skill may respond the following if there is an error in the skill or has blank response -

“There was a problem with the requested skill's response” or “undefined” as error response.


Good example:

User: Alexa ask engine sounds to play car sound

Alexa: I will play the car sound.

<Skill plays car sound>

Bad example:

User: Alexa, ask engine sounds to play the Car sound.

Skill: I will play the Car sound.

<Skill plays the bark sound>

Please Note - The above bad example clearly shows that the user asks the skill “engine sounds” to play “car sounds” but instead the skill plays plays a “bark sound”. It is important to have the skill give correct and appropriate response when a user ask for a specific function or information. Incorrect or irrelevant response might lead to poor customer experience and also failure of certification.

To understand more about the errors and issues with this test case, please follow the below tech documentation for additional insight into resolving problems with contextualized TTS or audio playback response.

text-to-speechcertification test caseaudio playbacktts
1592542847646.png (88.4 KiB)
10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.



Taira@Amazo contributed to this article harshit contributed to this article