July 14, 2016Zoey Collier
Craig Johnson, president of Emerson’s Residential Solutions business, claims it was inevitable. “Thermostats are no longer just passive HVAC controllers hanging on your wall. The convergence of wireless and mobile technologies allowed us to develop a thermostat that allows better temperature control, programmability and scheduling, as well as remote access.”
Even before Amazon’s Smart Home Skill API was publicly released, Johnson was excited about smart home. Prior to Smart Home, Emerson had a fully functional mobile app and internet portal our customers could use to control their Sensi thermostat remotely. But integration of Alexa is a natural extension of that remote access and remote functionality.”
In February 2016, Johnson’s software development manager, Joe Mahari, jumped on board the Smart Home beta program. In just four weeks’ time—and by the time Amazon officially launched the Smart Home Skill API—Mahari’s team had built and tested its Sensi Smart Home skill and passed certification.
The Smart Home Skill API converts a voice command, such as “Alexa, increase my first floor by 2 degrees,” to directives (JSON messages). The directive includes:
It then sends the directive to the methods implemented in the Sensi skill.
According to Mahari, Emerson implemented three main directives. Examples of these are:
The Emerson team agrees the skill and API were well packaged and supported, end-to-end. “Amazon defined the use case very crisply,” said Johnson. “We received a deck of scenarios to achieve, plus integrated logging, systems’ checks and documentation. These were essential to our success.”
Mahari says it was invaluable that the Amazon team connected with them daily. “For example, we had some concerns about how to increase or decrease the temperature during auto-schedules. But working directly with the Alexa team, we figured out how to make it work.”
So, if working with Amazon’s support and the API itself went so smoothly, what were some challenges the Emerson team faced over the four-week project?
“Emerson built their own automated test suite for performance and load testing, but they needed to make sure it held up in the real world,” said Mahari. “So we worked closely with AWS to spin up thousands of simulators, all sending real directives to us, so we could see how well we responded.”
That kind of testing support helped Mahari identify and resolve a couple of issues and easily go back and retest, ensuring Emerson’s skill would be able to handle thousands of devices.
Johnson says four weeks may sound like a short period, but “Certification was a challenge, because the technology and integration is so new. They tell you what they will test but not exactly how they’re going to test.”
The certification tests were still being refined during the Smart Home beta phase so “we didn’t make it over the goal line the first time,” Johnson said. The second time, they passed by understanding the pass-fail nature of the guidelines.
Mahari says most of his team use .NET and Visual Studio. Since a lot of the integration is Lambda AWS, they also use node.js. “We all have node.js installed on our machines so we can test our Lambda functions,” he says.
He strongly recommends developers also use a Node Package Manager tool, like Lambda Local. “This is a user acceptance test (UAT) framework that mimics Amazon’s AWS services locally, without having to always go out to the cloud.”
What would Emerson do differently? They say when in doubt about a use case, ask the users before implementing it.
“As an early technology partner, we made some implementation choices in raising or lowering the temperature set point. Using our existing app model, it made perfect sense to raise the set point, say, 2 degrees above the current ambient temperature. But that’s not how users saw it,” says Mahari. It turns out that Alexa users were more aware of current set point, not the ambient temperature.
“Interpreting the intent of a voice command is different than someone tapping on a thermostat or smartphone app. We frequently seek user feedback, and this reinforced that practice. If we’d asked that particular question of potential users first, we’d have saved some time and rework,” said Mahari. “Another lesson learned along the way.”
Johnson says Emerson has seen a 20 percent adoption of the Sensi Smart Home skill across new customers since release, but he says the best metrics are the corresponding boosts in product sales and customer satisfaction.
Emerson polls its customers frequently about the factors affecting their purchase decisions. Johnson says the numbers indicating customers see home automation as “important” or “very important” continue to climb month after month.
Mahari says, “We continue to see Echo and Alexa as providing more robust functionality and a better customer experience when compared to the other major competitors in this space. Whenever Amazon expands the kit’s functionality, taking that next integration step is definitely part of our plan.”
Share other innovative ways you’re using Alexa in your life. Tweet us @alexadevs with hashtag #AlexaDevStory.
Are you ready to build your first (or next) Alexa skill? Build a custom skill or use one of our easy tutorials to get started quickly.