4 Pro Tips for First-Time Alexa Skill Developers

Trish O'Connor

Last fall, I worked on three new voice experiences in partnership with Monumental Sports and EntertainmentThe Capitals SkillThe Capitals Flash Briefing, and The Wizards Flash Briefing!

You can ask Alexa about the Capitals’ score, the next game, your favorite player, the Caps’ record, and much more. Plus, you can get the latest news about roster changes, injury updates, and more team information using the Caps and Wizards flash briefings.

There are millions upon millions of Alexa devices in American homes. Voice experiences have a lot of potential to help brands interact with their users in a whole new way; so in preparation for the rise of devices (and branded voice experiences), here are four key takeaways for anyone developing an Alexa skill for a client.

Give yourself time to submit for Amazon certification.

Image Credit: ISL

It will take longer than you think. In brief, this is because Amazon manually tests the skill, so it takes time. We’re still in the early stages of Alexa development—tools and best practices are changing all the time. Occasionally the documentation is slightly disconnected from the latest SDK. When you are working up to a client’s deadline, you need to allow for the testing period and for potential feedback and resubmission. Give it at least a week— longer if it’s your first skill.

If you’re a developer looking for more details on the certification process, check out these handy tips and tricks.

Nail down the structure early.

The world of designing for voice experiences comes down to intents, utterances, and slot values. Using an example from The Capitals Skill, if a user asks “Are the Caps winning?”, what they really want to know is the score of the game. Hearing the score, then, is their intent (get it?). The utterance they use to express that are the words they speak. Each intent has many utterances. The user could say the words “What’s the score?” and have the same intent.

Let’s get a little more granular: slot values are variations in the utterances. There are multiple ways to refer to the Capitals hockey team – the Capitals, the Caps, Washington, DC, etc. In the utterance, “Are {team} winning?”, you can replace the slot {team} with any of the names for the Capitals and still mean the same thing.

That’s most of it— if you can identify the users’ desired end results (the intent) and the different ways they may express that (the utterances and slot values), then you have the basic structure for building an Alexa skill. The earlier you have that conceptual structure, the more time your developers have to make it a reality, and the more time you have to submit for certification.

Demo with your client early.

Once they actually hear Alexa speaking about their brand/products/services, they will have a million more ideas! This makes it fun as you tend to move from the basic, necessary intents to the more fun, creative ones. Again, this comes down to time – the earlier you get to this, the more time you have for fun intents. Be sure to budget time for several iterations and updates after client review.

Make an Amazon developer account for your client.

When your skill hits the Amazon store, you want to make sure your client’s name is right there with it. Right now, there’s no way to do that on your own account, so you’ll need to create an account for them and add your developers to the “Developer Role” on the client account.

It’s simple to set up and it will make your client very happy. Not only is their name attached to their skill, this will also give them access to the all-important metrics and data points – total unique customers, average number of utterances, total successful sessions, and total conversions— all without giving them access to the other skills on your account.

Author
Trish O'Connor
Trish O'Connor
Trish O’Connor is an Engineer at ISL (iStrategyLabs), where she builds interactive digital and physical experiences using exciting emerging technologies like AR, VR, and Amazon’s Alexa. She studied electrical engineering at Tufts University and we...
Trish O’Connor is an Engineer at ISL (iStrategyLabs), where she builds interactive digital and physical experiences using exciting emerging technologies like AR, VR, and Amazon’s Alexa. She studied electrical engineering at Tufts University and we...