Hear It from a Skill Builder: How to Create a Persona for Your Alexa Skill

Adva Levin Aug 15, 2018
Tips & Tools Tutorial

Editor’s Note: When creating an Alexa skill, it’s important to think about you skill’s personality. This helps you create authentic dialogs and more engaging experiences for customers. Today we welcome a community expert—Adva Levin, founder of Pretzel Labs and winner of the Alexa Skills Challenge: Kids—to share some best practices.

When designing a graphical user interface (GUI), we have many creative tools to design the look and feel we’re going for. We can use colors, fonts, images, animations, and text. When designing a voice user interface (VUI), our main tools are voice, word choice, and sound effects. One of the most powerful elements we can use to create an engaging experience for customers is a combination of these elements, which is called a “persona.”

A persona is part of someone’s character, as perceived by others. In a VUI, since we can’t see the person speaking, users paint a picture of the character in their minds based on the voice they’re hearing and the words the character is saying. The beautiful part about this is that a persona can change your user experience from an interaction to a relationship.

When conducting user testing, I’ve often heard people say things like, “Alexa, I love you” and “My Alexa is smarter than your Alexa.” They’re not saying these things about their computers or smartphones. They’re experiencing such strong feelings because the persona is so inherently defined in a voice experience that even though they’re talking to a small speaker, it has a similar emotional effect as talking to a person.

Creating a consistent and engaging persona in your skill can increase conversion, engagement, and the overall way users feel about your brand or product.

Here are some tips for creating a persona for your Alexa skill to deliver delightful voice experiences.

1. Pick the Emotional Reaction You’re Going For

Your persona should reflect your product’s voice (literally), and be geared at fostering the appropriate emotional reaction you want your customers to feel. For example, if you’re designing a financial skill, you may want to go for a persona that generates trust or credibility. If you’re designing something related to health and wellness, perhaps you want to create a sense of empathy. If you’re designing a game it could be excitement and fun.

2. Create a Sketch of Your Persona’s Characteristics

If your skill was a real-life physical service, who would you ideally hire to talk to your customers? What’s their gender, age, and background? Would they dress casually or business-like? How would they speak? Would they be very formal, or friendly with customers? This is not something you always have to communicate directly to your customers, but having an actual person in mind can help with writing a consistent character.

3. Choose the Right Voice for Your Persona

Choosing Alexa’s voice or a text-to-speech (TTS) voice from Amazon Polly lets you create dynamic content and be flexible when editing your speech. There are many voices to choose from to create different characters. Alternatively, you can record voice actors to create a more human feeling, which is great if your brand is already identified with a specific person. In this case, you’ll need to record the actors every time you’d like to add new content.

For example, Kids Court uses Amazon Polly voices to make customers feel like the judge is impartial and objective. It also allows us to add fresh content on a weekly basis. 

4. Write the Dialog. Every Word Matters!

Would your persona say sentences like, “Hello sir, how are you doing this evening?” or “Yo, what’s up?” Think of a few key phrases that are likely to appear in your skill and how your persona would say them. Try to be consistent with the character, yet write a variety of responses for every use case to create a more engaging experience.

5. Act Out Your Dialog

If you’re using a text-to-speech voice, play and edit it with SSML until it sounds great. Some words won’t work well in TTS, and some sentences that read wonderfully on paper don’t sound quite right when spoken out loud. Use short sentences, breaks, and pauses to create an experience that sounds more natural.

Example: Creating a Persona for a Restaurant Skill

Let’s say we’re designing a restaurant skill that takes food orders from customers. How can we give customers a sense of which restaurant they’ve arrived at and only with voice? The neutral choice would be using Alexa’s default voice and saying something like:

“Hello, this is Joe’s restaurant. We have some specials tonight. Would you like to make an order?”

Let’s try one more time, but with feeling. In the previous example, it could have practically been any restaurant. Now I want customers to feel like Joe’s restaurant is a friendly, casual diner. We can use a different voice, and change the wording to something like:

“Hey, welcome to Joe’s! Got some great specials today. What can I get ‘ya?”

See how that sentence has the exact same meaning, but it has a different tone? Now, if we were going for a more upscale place, we could use a formal word choice and perhaps a different voice:

“Good evening ma’am, welcome to Joe’s. We have a few special items on the menu tonight. What would you like to order this evening?”

See how these examples sound like different restaurants, although the meaning is essentially the same? For many skills, picking the right tone and persona could help increase engagement and create a better user experience.

More Resources to Build Engaging Alexa Skills

Follow Adva's tips to build a persona that can deepen engagement between your skill and your customers. For more tips on building highly engaging voice-first experiences, check out the following resources: