While smartphones represent the most convenient way for businesses to engage their customers, the emergence of IoT is expected to usher in a new generation of devices, ending the reign of smartphones. As mobile apps are no longer made just for smartphones, we believe the key shifts in mobility will center on how we interact with the device rather than on the changes to the device itself.
People are increasingly embracing intelligent apps that allow them to execute tasks quickly and easily with zero or low touch; with these apps, they also have relevant, contextual information at their fingertips that allow them to make informed decisions.
The intelligent app opportunity
How can businesses get ahead of the fast-moving trend towards intelligent apps?
When it comes to enterprise apps, we have an opportunity to provide relevant information at the point of decision-making or action. In the consumer space, we would be well-served by staying current on evolving user needs and expectations and responding to them in a timely way.
Four paths to intelligent apps
There’s usually more than one way to do something, and that goes for intelligent apps. Here are four ways to make apps smart:
- Intelligent interactions
- Intelligent APIs
- Intelligent data
- Intelligent voice
1) Intelligent interactions
What are Intelligent interactions and their benefits?
Broadly speaking, intelligent interactions occur when an app completes a task without users having to tap or click their way through a command resulting in a streamlined experience. There is a huge market opportunity to migrate from the legacy ‘touch-and-navigate’ user experience to the modern ‘converse, interact and accomplish’ paradigm.
Ways that intelligent interactions can take shape include the following:
- Apps perform tasks automatically by receiving cues from user intent signals – rather than through explicit user commands.
- Apps execute on voice commands, removing the need for users to tap and type.
- Rapid discovery of app content through new ways of app development like Android Instant Apps and Android Slices.
A prerequisite for building intelligent apps is a deep understanding of business processes. Building intelligent interactions starts with these questions:
- What is the purpose of the app?
- What are the important transactions associated with the app?
- How many ways are there to engage with the app?
- Where is the app targeted to run?
Examples of intelligent interactions
- Android instant apps: Users can try native Android apps without installing them. This technology can help to increase user engagement with the app and drive more installations.
- Android slices in Google search: UI template that is rich and dynamic, providing interactive content in the Google search app and Google Assistant.
2) Intelligent APIs
Traditionally, web services are exposed as REST or SOAP APIs over a remote server or in the cloud, and are always utilized by mobile apps over the network. Emergence of AR and VR are paving the way for modern mobile app development in areas like retail, manufacturing and sales.
What has changed over time to achieve new ways of building intelligent mobile apps?
- Mobile Native APIs: These are mobile SDKs and libraries that are running within the mobile device utilizing on-device CPU and GPU to deliver high performance and efficiency.
What benefits can intelligent mobile native APIs deliver?
- Generic data will give way to more relevant information.
- This will provide users with greater insights at the point of decision-making.
Examples of intelligent APIs – Google VR SDK and ARCore
- VR SDK: Provides native APIs for features such as input, controller support and rendering.
- ARCore: Enables AR app development with essential AR features like motion tracking, environmental understanding and light estimation.
- Apple AR Kit: In addition to common AR features, Apple’s AR Kit also provides relevant and shared AR experiences.
3) Intelligent data
According to an Ericsson study, 70 percent of the world will be using smartphones by 2020. As a result, data collected by mobile apps and sent to servers for deriving outcomes will become exponentially massive.
How does this affect the development of intelligent mobile apps?
- Increased wait times for deriving outcomes and recommendations.
- Higher bandwidth utilization by mobile apps.
How can this be rectified?
- Process data within the mobile device.
- Transfer only high-level data to the server.
Examples of intelligent data
TensorFlow for mobile: Traditionally, the high-end GPU computing required for deep learning and neural networks have always occurred at remote cloud servers. Yet, sending and receiving all this big data over the network is both expensive and time-consuming. TensorFlow for mobile makes it possible to build interactive intelligent apps.
Core ML 2: Built on top of low-level technologies like Metal and Accelerate, Apple’s Core ML 2 provides a way to run machine learning models on the device so that data doesn’t need to leave the device that is to be analyzed.
Additionally, developers should enable online consumers to protect their privacy by:
- Obtaining consent from user before starting to collect the data.
- Adhering to General Data Protection Regulation (GDPR) guidelines on data privacy.
4) Intelligent voice
A recent statistic published by VoiceBot AI states that more than half of U.S. smartphone owners use voice assistants. This trend underlies demand by consumers for mobile apps to handle voice-based commands so that users can take actions quickly.
Benefits of voice-enabled mobile apps include:
- Greater user engagement;
- Improved accessibility for the visually impaired.
Intelligent voice examples
- Google Voice Actions: Voice requests in Android can directly reach mobile apps, allowing users to quickly and easily complete any task.
- SiriKit: Besides deep linking to an app, developers can use Siri to intelligently pair users’ daily routines with their apps to suggest convenient shortcuts when needed.
Futuristic use cases include:
- Voice biometrics within mobile apps.
- Inferring emotions and context from voice.
- Identifying and handling multi-user voice mode.
