{"id":152,"date":"2016-03-31T10:06:47","date_gmt":"2016-03-31T10:06:47","guid":{"rendered":"https:\/\/portfolio.cherylplatz.com\/?p=152"},"modified":"2022-07-31T01:21:45","modified_gmt":"2022-07-31T01:21:45","slug":"alexa-do-not-disturb","status":"publish","type":"post","link":"https:\/\/portfolio.cherylplatz.com\/?p=152","title":{"rendered":"Alexa: Do Not Disturb"},"content":{"rendered":"

In parallel with our efforts on Alexa Notifications, I was asked to drive the completion of the Do Not Disturb feature. DND applied not just to Notifications but to Communications (calls\/messaging), which meant a separate team of stakeholders and faster deadlines than the longer-term Notifications project.<\/p>\n

As with Alexa Notifications, I led the design side of this effort in partnership with Carl Mekala (and Sharbani Roy) from the product management team.<\/p>\n

This was my most “traditional” VUI design engagement during my time on the Alexa team – while some was new-product (Echo Look) and some was more systems design (Notifications\/Interrupt Model), this work was a fairly representative VUI design process. Do Not Disturb shipped with the Echo Look and the launch of Alexa Calling in Spring 2017.<\/p>\n

Design Process<\/h1>\n

For our design process on Do Not Disturb, I started by pulling existing transcripts from Alexa users – and from lab tests – where our transcriptionists had marked a notification intent. (These customers usually didn’t know the feature wasn’t available yet, and were simply trying to get smart home or app notifications from Alexa.) That provided a basis for the mental model customers were likely to apply, and a start for our sample utterances.<\/p>\n

Our initial discussions also involved the app team – what intents must be voice-enabled, which intents should be app-enabled, and which intents should be available in both places? Our initial focus was recurrence via the app, and one-time DND requests via voice.<\/p>\n

Audio Mocks & Peer Critique<\/h3>\n

After a round of level-setting with product management, we set an initial group of intents. Based on the kickoff discussion and my own brainstorms, I did audio mocks of several approaches to the prompts and structure. These prompts were brought to my design peers and stakeholders for feedback via a quick email survey (a common practice on our team given the speed of iteration and our split locations.)<\/p>\n

Early Audio Mocks: Prompts-DND<\/a><\/p>\n

At the same time, I reached out to our audio design team with the hypothesis that a sound effect (“earcon”) might be preferable to a full spoken prompt for starting DND in certain cases, or perhaps could be used for recurring DND. This partnership resulted in a pair of sounds: DND start\/stop. In the end, we decided to use the sound effects as a gate: they play after any VUI prompt when enabling DND, and before any reply when turning DND off. <\/p>\n

Storyboarding and Sample Dialogs<\/h3>\n

As feedback from our peer critiques and stakeholders began to roll in, I pulled together a series of rough storyboards to begin fleshing out our end-to-end story. This storyboarding work paralleled our work on the sample dialogs for the Do Not Disturb feature. Below are a few example frames from one of my earlier DND storyboards. In the end, we streamlined our dialogs quite a bit – choosing to omit qualifications like “on this device” from the system’s responses.<\/p>\n