AI Helps Fuel Rush to Real-Time Data Platforms, ISG Says

By 2027, more than three-quarters of enterprises will be able to process events and streaming data in real time, making them more responsive to customer needs and other requirements, according to new research from global AI-centered technology research and advisory firm Information Services Group (ISG) (Nasdaq: III).

The ISG Buyers Guides for Real-Time Data, produced by ISG Software Research, provide rankings and ratings of over 25 software providers and their products for supporting the architectural shifts required for data to support AI for operations. The research finds that companies are seeking the ability to process and analyze data in real time to enable applications that deliver AI-driven personalization and recommendations. While most organizations have long relied on batch data processing, which was necessary when computing power was more limited, they are now looking for a new approach that supports the full benefits of AI. Real-time machine learning on operational data delivers instant, relevant information for faster decision-making.

“Real-time data processing lets enterprises operate at the speed of business, acting on events as they happen,” said Matt Aslett, director of research, ISG Software Research, and lead author of the reports. “As consumers seek AI-driven interactive applications, companies need information architectures that include streaming data and event processing.”

Until recently, real-time data processing was limited to industries with extreme high-performance requirements, such as financial services and telecommunications. Only 22 percent of enterprises analyze data in real time today, ISG says. However, the core concepts and technologies are proven, mature and readily available. The new ISG research finds that the top software providers have matured to deliver nearly 80 percent of the overall product and customer experience required to support enterprise needs.

Real-time data processing begins with an event, which may be any change of state, such as a sensor identifying a new temperature reading, the reports say. With effective application integration, sensors, devices and applications can share messages about events as they occur, increasingly via application programming interfaces. Real-time data systems ingest, filter and aggregate data about events through stream processing. This forms the basis of stream analytics, which analyzes event data using SQL queries, machine learning or inferencing based on generative AI models.