Businesses are dealing with three tectonic forces that are now challenging IT everywhere – digital transformation, channel proliferation and real-time AI-decision management. While both digital transformation and channel proliferation have been around for some time, real-time AI decision management is playing havoc with this shifting landscape. Businesses need new solutions that can meet the unrelenting demand for fast, accurate digital services, and consumers are not going to give them any slack as IT searches for new ways to solve these challenges. 

These are issues that affect enterprises and smaller concerns, and everyone in between. Whether the systems are developed in-house, running on the cloud, are using .Net on SQL Server, or accessed on mobile devices, these apps place huge amounts of workload on all components. Many of these components, such as the backend databases, were not designed for the speed and volume that businesses require to be able to provide a competitive advantage. 

Digital transformation

Let’s examine each of these forces, beginning with digital transformation. In a recent webinar where Onepoint, an IT consultancy in the UK, joined GigaSpaces, Dr. Madassar Manzoor, a solution architect, discussed a bank they have been working with for a number of years. The bank has over 650 branches and serves 15 million customers. Whereas in 2004 they processed 10 million simple transactions per day, twenty years later they are processing 200 million transactions per day via internet and mobile banking. These transactions comprise complex calculations, but the 15+ core banking systems that process these transactions, such as loans and credit cards, have been developed in .NET and use SQL Server. None of these systems was designed to handle millions of concurrent hits by customers every day, and these systems are now pushed to the limit of their capabilities. 

Channel proliferation

As consumers demand more and more self-service capabilities – especially on mobile devices, the workflows that support these decision trees have grown much more complex. A business’s call center now usually offers a bot and recently perhaps WhatsApp, in addition to its online and mobile applications. All of these channels place heavy demand on operational systems that weren’t designed to support this usage, let alone support the machine learning interactions that enable higher level interactions and decisions. To further complicate this issue, these back-end systems were not designed for the direct access that apps offer to consumers. These systems were created for the use of in-house staff (not bots) who would respond to calls, not for the high volume of queries, 24/7, that strain the ability of these systems to keep up with the demand. 

Real-time AI-decision management 

Many businesses have moved from phone and email interactions to consumer portals. Even non-consumer industries such as insurance, can offer their users the ability to download applications and upload contracts, and even return a decision about a policy without human intervention. Whereas staff first began making decisions based on simple business logic, businesses have since adopted predictive models that can adjust to changes in customer behavior and to changes in the market. Predictive models are used for many business decisions, such as to upsell, detect fraud and to provide a risk score. GenAI is taking this even further and is able to resolve issues on its own, instead of directing consumers to specific web pages or to a human staff member. 

With this perfect storm of more services in the hands of customers, the interactions happening via several, often concurrent channels, compounded by the increased complexity of these operations. The steep increase in the number of decisions required at the point of contact with the customer, is at odds with the number of decisions required by these services. 

Real Time Data Processing with a Data Integration Hub

Fortunately, a proven solution has come to the rescue of legacy infrastructure that continues to serve modern digital applications. This involves continuously bringing the necessary data (not the entire data warehouse) for in-memory processing, at the frequency at which the data is created. This real time data processing of events and transactions occurs in a data integration hub, a highly scalable solution that centralizes and consolidates the data. 

For more information about how an off the shelf Hub such as GigaSpaces Smart DIH has proven so valuable for retailers, banks, insurers and other businesses, click here to view the full webinar. Smart DIH’s event driven architecture offers powerful compute capabilities, high availability and scale.