In this blog post Exploding Phone's Business Development Director, David Stephenson, considers the future for Smart Mobile Apps:

The next generation of ‘Smart’ apps will dramatically improve and extend the mobile experience. Apps will achieve this by learning from patterns of user behaviour and thereby enhancing the app’s deliverables. Artificial Intelligence techniques, for example Machine Learning, in combination with advanced device sensors, location-based capabilities, and connectivity of today's and future generations of portable devices will increase utility for app users and create competitive advantage for businesses.

Machine Learning enables a app to be ‘trained’ to perform a task by sourcing data and for the program to find and refine patterns and connections in the data over time. Once the app has been trained it process new data to make deliver seemingly intelligent decisions. In the past Machine Learning has required significant computational power, however, advances in mobile processing power has enabled mobile apps to adapt and deliver information that mimics intelligence, i.e., sense, perceive, learn from, and respond to the environment and their users. This trend will be augmented by ubiquitous access high-speed access to online computational resources, but mobile apps themselves should also perform AI tasks locally if they are to provide a consistent user experience.

For organisations that wish to exploit smart mobile apps, Machine Learning allows apps to harness dynamic business analysis techniques, data mining models and adaptive rules to simulate and refine pseudo-cognitive decision making. This has the potential to improve the efficiency and effectiveness of how businesses engage with customers, employees, partners, the community and potentially the wider environment. To ensure the best possible adoption of Artificial Intelligence when designing and building mobile apps the solution should strike a balance between accessing and processing the largest possible data sets while at the same time being frugal with processing and communication requirements. Harvesting data can be enhanced via a deep integration between the app and the host devices sensors, etc., while the efficient use of information is helped by a detailed understanding of data science and AI algorithms. Crucially the code created needs to undergo rigorous testing to ensure resilience and the quality of outputs.

Many apps are already using Machine Learning to deliver personalization, recommendation services, predictive searches, improved security, and analysis of personal metrics such as health and fitness data. Notable examples of apps already using AI include Snapcha’s recognition and manipulation of facial features and Netflix and Tinder apps attempting to understand and anticipate personal preferences. This trend will undoubtedly continue with apps that exploit AI disrupting markets and industries by making previous app paradigms redundant. However, the effectiveness of apps incorporating Machine Learning/AI techniques is constrained by both the data and the model used to process the data. As such it is essential to partner with an experienced app development team with a combination of traditional app development along with AI and Data Science skills.