Integration as a machine learning facilitator
AI (artificial intelligence) is quite a trendy concept with seemingly endless possibilities. AI develops the more it is used, and machine learning needs very large quantities of data. Data masses can easily be gathered by integrating incoming data streams. For example, a gasoline company can gather data about refueling and the number of transactions. Sales data are stored in a data warehouse and then used to analyze sales. Based on this analysis, fuel prices can be automatically set to the optimal level. “Optimal” means the price that achieves the largest fuel sales at the highest price people are willing to pay at any given time.
Help for employees on the move
Process automation, interaction and IoT (Internet of Things) can also help people who work in various locations. A tachograph installed in a car sends trip data to a central computer via the interface provided by the integration solution, and from there the data can be sent to the payroll department and other places. There, process automation is used to help calculate travel compensation forms, per diems and mileage according to the distances driven and the location data. The car itself can also send data to a central computer. With the help of IoT, service can be booked automatically based on notifications sent by the car, either after a certain number of kilometers or when a fault appears.
Shopping delivered straight to your trunk
IoT is expanding the Internet to devices and machines that can be controlled, measured, and equipped with sensors across the Web. HiQ developed an interesting IoT service concept for Volvo: you shop online, the store collects your items, and then delivers them to the trunk of your car. And of course, in such a way that the car doesn’t need to be in a certain place at a certain time — instead, the delivery finds your car using location data. The person delivering the shopping opens your trunk using a single-use code, meaning the car owner doesn’t need to be there at the time. The delivery status can be tracked using a Volvo service portal and the customer gets a notification of the completed delivery to his or her smartphone or by email. At present, the service is operating with certain vendors and Volvo cars in Sweden and Switzerland.
Mini-services: rapidly published content from existing systems
SOA (Service Oriented Architecture), developed 20 years ago, was targeted at the provision of the most popular commercial services. The service interfaces made in one project could be used in the next project. However, the problem with these “macro-services” was that they excessively aimed at general usability, and the service development was slow and expensive. Moreover, services aged quickly as business developed. In the past few years, micro-service architecture (MSA) has increased in popularity. Micro-services are small services that run entirely independently. They can be adopted independently of other services or systems. The problem with micro-services is that to be completely independent of other data systems, each service must have its own data warehouse. Thus, it cannot rely on other systems’ data (such as CRM). Mini-services came into being in the space between these two. A mini-service is like a micro-service, except it has no requirement for complete independence. It can be loosely coupled with other systems and interfaces. Any data or service that is requested from a back office can be offered to other parties as a mini-service. A mini-service could be, for example, an interface for placing orders, or a request made by a mobile app to a back office for an account balance. The service can be installed and scaled independently. This type of service is used a lot when people build new services quickly on the back of existing ones. Gartner thinks that mini-services will be the next big architecture hit in the coming years.
Open interfaces and API management
Banks’ interfaces have been opened through legislation: the PSD2 directive means that banks must offer third parties and other operators in the banking sector the opportunity to use customers’ data (with their consent) to offer new services. The same philosophy regarding open interfaces could also be applied to retail, for example. A vendor could link a mobile device to the store’s product database, and a system installed in the device could place product restocking orders automatically. When the data obtained via API (application programming interface) shows that stocks of strawberry jam are running low on the store’s shelves, the mobile app talks to the back office and orders more. Sharing, managing and securing APIs is simple with the help of an integration platform.
Benefits for vendor and buyer
Process automation and the services that can be connected to it will offer more in the future: not just more savings for the companies that take advantage of them, but also easier, more personalized, and faster service than ever before. Not to mention quality — which is what customers appreciate.