Designing Mobile Apps using the Speech API
Space is at a premium on mobile apps. You can’t simply shrink the full functionality of your web or desktop apps to fit into a mobile screen and expect to provide a compelling user experience. You need to rethink how you obtain and disseminate information.
Most complex tasks should be reimagined to be simpler and streamlined. On the desktop, you could gather information using sophisticated “wizards” that guide a person through the process of providing information. Mobile devices are far more task-oriented and specific, and because screen real estate is precious, it’s difficult to provide visual cues or breadcrumbs along the way to keep someone informed about where they are in the process. In short, wizards don’t work on mobile devices for obtaining complex data. Moreover, people hate inputting data with soft keyboards, so complex forms also don’t work very well.
When people are presented with complex data entry tasks in apps, they will typically abandon the app rather than continue to use it.
Instead, consider using speech recognition for some data entry tasks. Audio prompts followed by speech recognition tuned for specific grammars will enable you to obtain user input, while respecting the limitations presented by mobile devices. With the AT&T Speech API, you get a proven speech service that’s easy to use and incorporate into your apps, and which will help you build more engaging user experiences.
Here are five suggestions for including speech recognition into your apps:
- Don’t go overboard. Using the AT&T Speech API is incredibly simple. So simple, in fact, that you may be tempted to provide it as an option everywhere. This is never a good idea. Use it judiciously in areas where user input is difficult via the soft keyboard and the impact will be high on your app.
- Re-think how you do user input. In line with tip #1, and not strictly advice for speech recognition, is experimenting with different styles for user input. Instead of a complex set of strings, can you make do with a simple button or toggle switch? Can several parts of your app add up to getting all of the information you need? Rethink everything about design when targeting mobile devices, and consider speech a key part of that process.
- Use good visual cues. Be smart about how you prompt the user for voice input. Think about adding a “What you can say” screen as part of the “new user experience” for your app. Gradually introduce users to voice input for your app.
- Use pre-built grammars. You can certainly use a free-form grammar in places where it makes sense. But if you’re asking people to specify their city or state, using a geography grammar will result in far greater accuracy. Similarly, a phone number grammar will make entering phone numbers more reliable.
- Take the time to build custom grammars. Custom grammars for domain-specific apps will help you increase the accuracy of your apps, and obtain user input quickly. Say you’re building a prescription tracker app for doctors. Building a custom grammar with names of drugs will make it easier to recognize input. Custom grammars involve a considerable amount of work, but the benefits to your users will be very high, which will result in greater and continued engagement with your app.
Speech recognition support in apps can be a critical component in bringing sophisticated data entry to mobile devices. Using speech judiciously and in a well-planned out manner can result in greater satisfaction and engagement with your app.