Troubleshoot Your Alexa Conversations Skill

Having trouble creating a skill with Alexa Conversations? The following sections show possible resolutions for common issues. The sections cover issues for editing the model, building the model, responses, slot types, utterance sets, launching the skill, testing the skill, and certifying the skill.

You can also use the following support channels:

For a list of Alexa Conversations known issues, see Known Issues.

Editing the model

Issue: Edits to my model are not saved

Symptoms
As I edit my Alexa Conversations model and switch between tabs, my data is not saved.

Try this

  • Save often. Navigating between tabs doesn't always count as activity for extending your developer console session.

  • Be sure that you only use one browser tab to edit your model from the Build page. It's ok to have multiple browser tabs open — for example, one tab for the Build page, one tab for the Code page, and one tab for the Test page — but if you have two Build tabs open and you make changes in one of the tabs, you're at risk of losing data.

Issue: The right panel indicates that a variable is not defined

The right panel displays the error message "Variable is not defined".

Try this
Be aware that this error occurs if you define a variable during dialog annotation, use that variable in an utterance set, API definition, or response, and then delete the variable.

Building the model

Issue: I get errors when I build the model

Symptoms
When I try to build the model, I get Build Failed messages.

Try this

  • Check the Build Failed messages for any information or line number that you can use to examine your annotated dialogs. For details about errors, see Alexa Conversations Errors.

  • Check your Alexa Conversations dialogs to make sure that all the user lines and Alexa lines have the success icon (green circle with checkmark). If a line has the failure icon (exclamation mark), the line isn't fully configured. Examine the configuration on the User Input or Alexa Response slide-out panel to find the missing elements. Missing configurations are a common error because you sometimes must click + or Create after entering or selecting options.

  • Make sure that you have at least one dialog that has at least one of the following:
    • A response with an argument
    • An API with a required argument and a request template
    • An API with an optional argument and an opening sentence informing that argument
  • Check Work with Dialog Acts in Alexa Conversations to verify that your dialog act flow is supported.

  • If the error is MissingRequestTemplateV2, MissingDialogAnnotation, MismatchArgumentSize, MissingSampleUtterance, or SchemaError, see Common errors for suggestions.

Issue: I get a high traffic congestion message when I build the model

Symptoms
When I try to build the model, I get a high traffic congestion message.

Try this
Try again after some time or reach out to us on the Alexa Community Slack Channel (sign in or join) or the Alexa Skills Kit developer forum.

Issue: I get error code MissingSampleUtterance when I build the model

Symptoms
When I build the model, I get error message Missing sample utterance. At least one sample utterance is required. Error code: MissingSampleUtterance.

Try this
Make sure that the interaction model has at least one intent with a sample utterance. To access the intents in the developer console, expand Interaction Model > Intents in the left pane. For example, you can add a sample utterance to a built-in intent that your skill uses.

Issue: I get error code MissingRequestTemplateV2 when I build the model

Symptoms
When I build the model, I get an error message such as: Annotation of Request for argument "day" with slot type "AMAZON.DATE" to a Response is missing from the sample dialogs and must be added. Error code: MissingRequestTemplateV2.

Try this
Make sure that each required API argument has at least one Alexa turn that requests it. Also make sure that each API has at least one dialog that uses the Request Args dialog act. For examples of troubleshooting this error, see MissingRequestTemplateV2.

Issue: I get a build error such as MismatchArgumentSize after I update the model

Symptoms
After I make a change to my model, I get a build error that seems to apply to a previous state of the model before the change. For example, when I change an API definition from having two arguments to one argument and then rebuild the model, I get the error message Dialog Name: "<Dialog name>" - "<API definition>" only has 1 arguments in its definition, but 2 variables are assigned to it. Error code: MismatchArgumentSize even though there are no longer two variables assigned to it.

Try this
Delete the item (for example, the "API to Invoke" in the previous case) from the response annotation, and then add it back in. You might need to take a similar approach for utterance sets also. For more details, see MismatchArgumentSize.

Issue: I get error code MissingListRequestTemplate when I build the model

Symptoms
When I build the model, I get an error message such as: An Alexa Response with a Request Dialog Act for type "List<AMAZON.City>" is missing. Error code: MissingListRequestTemplate.

Try this
If this error appeared after you added an argument to your API to Request, check to see if the developer console automatically added a tag for the new argument in the API Arguments to Request field of the Alexa Response panel. If so, remove the new tag.

Responses

Issue: When I add a response prompt to my skill, the prompt content disappears after I save the skill

Symptoms
On a response page, I fill in Audio Response Name, add a prompt, and then click Update Prompt. However, when I save the model, the content of the prompt disappears.

Try this
After you add the prompt (but before you click Create Prompt), make sure to click + to the right of the prompt. Otherwise, the developer console doesn't save the prompt.

Issue: The responses from my skill don't use my custom response templates

Symptoms
Alexa doesn't use your custom response templates when responding to the user.

Try this
Verify that you enabled the Alexa Presentation Interface for your skill in the Interfaces section of the developer console.

Issue: Problem playing audio files in responses

Symptoms
Audio files that your response templates reference don't work.

Try this
Make sure that your audio files meet the audio format requirements for APLA. For details, see Audio file support.

Issue: When I map variables to arguments in a response, I can't find the variable I need

Symptoms
When I try to configure the variable-to-argument mapping for an API associated with an Alexa response, the drop-down menu doesn't provide the variable I need. I'm not able to manually enter a variable name either.

Try this
Select the API to invoke before you configure the response. Selecting the API to invoke populates the response variables with all the available options.

Issue: I can't create a new response during dialog annotation

Symptoms
I keep finding myself annotating a dialog and realizing I haven't created an Alexa response for a particular dialog line. As a result, I have to go to the Responses part of the developer console and completely re-enter all the same information to create a prompt.

Try this
You can create a response directly from the dialog annotation tool. When you use the Alexa Response slide-out panel, click the drop-down menu under Response, and scroll to the top of the list of options. Click Create New Response, the first option.

Issue: When I pass a custom slot type with values to an APL response, I get an error invoking the skill

Symptoms
When I pass a custom slot type with values to an APL response, I get an error and the skill won't invoke.

Try this
Alexa Conversations currently can't pass custom slot types with values to APL responses, so you need to pass the complex type (slot type with properties) to the response. To do so, return all the data from your API and then send it to the response. Then, reference that data in your APL. For example, if you return the following from the API:

return {
  apiResponse: {
    'key1' : 'val1',
    'key2' : 'val2'
  }
}

…plug the following into your response:

invoke api(..) => res => Alexa Response (res: Result) => APL|APLA

Then, in APL and APLA, do the following:

${payload.res.key1} and ${payload.res.key2}

However, if you only use a voice response template (APLA), this workaround is not required because you can pass custom slot types with values to APLA.

Slot types

Issue: I can't create a custom slot type with a list of allowed values

Symptoms
I want to create a custom slot type that specifies its allowed values in a list. For example, my slot type is called PizzaToppings and I want the values in the list to be green peppers, mushrooms, pineapple, and so on. The user interface doesn't let me create custom properties when I create the new slot type.

Try this
When you add the slot type, choose Create a custom slot type with values.

Utterance sets

Issue: I get the error "There is no Utterance Set with Dialog Act Invoke APIs for this API Definition."

Symptoms
I get the error "There is no Utterance Set with Dialog Act Invoke APIs for this API Definition. Error code: MissingOpeningSentenceForApi" even though I have an utterance set that uses the Invoke APIs dialog act for the specified API definition.

Try this
Make sure that your utterance set contains less than or equal the number of arguments the API definition contains. Otherwise, Alexa Conversations might find it unlikely that the utterance set is the invoking utterance for the API, because the utterance set provides more than the number of necessary input arguments.

Launching the skill

Issue: My skill receives the launch request

Symptoms
Your skill endpoint receives a LaunchRequest. If you selected Alexa Conversations as the default dialog manager, Alexa Conversations should receive the launch request, not your skill.

Try this
Confirm that you set Alexa Conversations as the default dialog manager for your skill in the Interfaces section of the developer console.

Issue: I get an error when I invoke my skill in the simulator

Symptoms
When I invoke my skill in the simulator, I get a red exclamation (!) message.

Try this
Verify that you specified an endpoint (that is, the location of your skill code). In the Build tab of the developer console, on the left, select Endpoint.

Testing the skill

Issue: When I test my skill using the Alexa simulator, the Alexa simulator displays <Audio only response> and the skill session closes

Symptoms
When I test my skill and type a user utterance and the Alexa simulator, the Alexa simulator displays <Audio only response>, and the skill session closes.

Try this
The interaction model must have at least one custom intent with at least one sample utterance.

Issue: My APIs have list arguments, and during runtime testing, Alexa says "Sorry, something went wrong"

Symptoms
Your skill has APIs that use list arguments, and during runtime testing, Alexa says "Sorry, something went wrong."

Try this
Make sure that your APIs don't take multiple list arguments. (Each API can have at most one list argument.)

Certifying the skill

Issue: My skill fails certification due to example phrase testing

Symptoms
My skill failed certification with the reason "ensure that example phrases returns the expected response."

Try this
The example phrases the certification team refers to are the user-facing example phrases entered in the Distribution tab of the developer console. Make sure that the user-facing example phrases have a matching utterance in the skill utterance sets and that these utterances return an acceptable response.