Step by Step

  1. Navigate to the Alexa developer console in a new tab alexa

  2. To save time, we’ve pre-created an empty Alexa skill called daily deal team # that we can start to populate with the following:

    • Invocation: This is what the customer will say to begin their interaction with your Alexa skill.

      • Enter the invocation as daily deal team #Click to copy alexa

    • Intent: This is an action that fulfills the customer’s request based on provided inputs.

      • Select Add Intent
      • Enter SurpriseMeIntentClick to copy alexa alexa
      • Select Create custom intent. On the next page, enter what you would expect the customer to say to invoke the intent. You can add multiple utterances, such as:
        • Surprise meClick to copy
        • Send me somethingClick to copy
        • surprise me its {name}Click to copy
        • Its {name} surprise meClick to copy alexa

    • Slot: You may have noticed in the above examples that name is in curly brackets. This is to indicate that this is a variable value as we’re going to need to know the customer’s name. We can either get that information from their initial utterance, or we can ask them for it.

      • Scroll to the bottom and observe the added Intent slots called name
      • Set the Slot Type to firstnameClick to copy to ensure the Alexa skill recognises the type of data being provided
      • Select Edit Dialog
      • Enable Is this slot required to fulfil the intent
      • Under Alexa speech prompts - type What is your name Click to copy and click +
      • Under User utterances - add My name is {name}Click to copy and click +

      This ensures your skill is able to collect the user’s name if they haven’t already provided it. alexa alexa

    • Navigate to the Validation tab

      • Select Accept only Slot Type’s values and synonyms - Click +
      • In the text box below - type sorry I didnt catch that whats your nameClick to copy then press + alexa

    • Scroll back to the top and select Build Model.

We’ve now created the basic construct of our Chabot that’s going to interact with the customer and collect information required for us to determine a suitable product and deliver it to them. Let’s now glue everything together before finishing our Amazon Personalize machine learning model.