As you no doubt have experienced, a new market of conversational devices and services has emerged. From Amazon Alexa and Google Home, to Facebook Messenger bot and intelligent assistants, we’re surrounded by conversation. (Between us, when was the last time you checked the weather on a device other than Alexa or Google Home?)
While those devices and services are new and extremely dynamic, we can find a few success cases. Amazon Alexa, Google Home, and Apple’s Siri are three worthwhile examples to learn from.
Below are some important factors that made these devices so successful and how they can help you build your own killer conversational application:
1. Find the direct path to initial success
Voice control bots and conversational UI services, in general, are actually not that new in our world. There were a few attempts in the past; even the old Nokia dumbphones had voice control features. However, most of these trials didn’t succeed. With zero patience, after one or two unsuccessful tries, users just moved on to the next thing or fell back to the web or mobile apps.
So, how do we find the direct path to success?
Help your users ask the right questions. This sounds pretty obvious, but it is actually crucial to the success of your Alexa skill or Google Home action. I learned that when I initially set up my Amazon Echo device at home. The mobile app directed me to ask Alexa specific questions for which she had good answers, such as “Alexa, what is the time?” or “Alexa, what is the weather today?” I immediately received correct answers and therefore wasn’t discouraged by a default response of: “Sorry, I don’t have an answer to that question.”
2. Think like a search engine, not a command line
Conversational devices, command line interfaces, and search engines have much in common. In all three, you provide the machine with a textual question and receive in return, a textual answer. But think about the user experience of command line versus a search engine. Whereas a command line will produce an error message if you don’t use the exact correct format, search engines will always return a result, whether it’s the answer to your question or a suggestion for a different search.
Here’s an example: to play music with Google Home, you could say: “OK Google, please play Taylor Swift station from Pandora.” In this case, Google will reach out to Pandora and search for Taylor Swift’s songs to play. But you can also use a shorter version for your request, such as: “Hey Google, please play Taylor Swift songs.” In the second version, Google will pick a music service and search for a Taylor Swift station there to play. If you just say: “Hey Google, please play music,” Google will search for your enabled music service and play one of your favorite stations. Acting like a search engine provides you and the user with much more options, as there are answers for all different cases regardless of the format and data the user supplies.
3. Offer real value and functionality
Finding your minimum viable product (MVP) is crucial for the success of any voice-control application, but beware of burning users by offering a limited product that does not provide value. Start with simpler types of features, such as content consumption. It can be listening to music or hearing the news, but also receiving information on your bank account balance, finding the price of an item, or even checking your next meeting on your CRM. Next, develop your secondary types of features that would let you create content or modify settings. This can be, for example, creating a new music station, transferring money between accounts, or adding a new meeting to your CRM. By providing a full-service application, you make sure your user sticks to your conversational application and doesn’t fall back to your web or mobile app.
4. Build an omnichannel service
Remember, new devices are not replacing old devices — they are only adding to the big basket of channels you must support. What Amazon Alexa started was instantly followed by Google and now even Apple. Users today want to get your services anywhere and anytime. To do that, you must provide a similar level of services on all the different channels. For instance, you might see a spike in requests during the early morning or late night coming from home devices, such as Amazon Echo and Google Home. However, during the day, you will receive more activities from Facebook Messenger or your smartphone intelligent assistant. Similar questions must get similar answers on all devices, no matter how your user is consuming your services.
5. Stay up-to-date with trends
Alexa, Google Home, and all the other voice-enabled devices are a new field of computing, and also extremely dynamic. As such, today’s top trends might grow huge, but they can also become obsolete and outdated in just a few months. This is why you must make sure that your applications stay up-to-date with users’ expectations. Some examples include: supporting new types of enabled conversations, adding new functionalities, implementing relevant UI and UX, and of course, supporting all the latest trendy devices.
6. Start now
Gartner predicts that by 2018, 30 percent of our interactions with technology will be through conversations with smart machines. The digital voice-activated devices market is expected to double itself every year by 2020.
Developing a voice-enabled application requires profound knowledge in machine learning, voice recognition, and natural language processing. In addition to that, developers must be extremely dynamic and flexible.
If you want to make sure that you get a bite of this cake, you better start now. Learn the market, find the value you can provide using voice-controlled and conversational capabilities, and start building prototypes that will enable you to grow with the market.