Tapping into Data Capital with AI and Machine Learning

As companies begin to understand the vast potential of their data, the question they face is: How does our business make the most of it? The answer lies in getting real-time insights that enable better business decisions and accelerated product development.

But what if the insights were used not just by humans, but by the systems themselves, leading to ongoing optimization at previously inconceivable speed and accuracy? That’s the promise of adaptive intelligence (AI) and machine learning, which are already impacting consumer experience with personalized shopping, self-driving vehicles, online wealth management, and virtual assistants. Here, we’ll look at how data is driving the coming AI revolution.

In its most basic sense, human intelligence is the ability to make decisions based on observation and experience. Artificial intelligence, then, is the same capacity in machines, and potentially as vast and complex.

Here at Oracle, we prefer to focus on a subset of AI called adaptive intelligence. Adaptively intelligent apps consume streams of raw data from multiple sources—such as customer experience, enterprise resource planning, supply chain management, and human resources—to develop predictive analytics. This is an ongoing and self-correcting process, so that these analytics are continually refined and adjusted to reflect changing input.   The key, of course, is high-quality data—and lots of it. “Data is the fuel that drives organizations towards automation,” Rich Clayton, vice president of Oracle’s Business Analytics Product, recently wrote. “But most organizations lack a comprehensive data strategy, one that seeks to acquire, curate, combine, and commercialize it.”

Chuck Hollis, senior vice president for Oracle’s Converged Infrastructure group, uses the example of hiring to explain how machine learning and adaptive intelligence work. In this case, the enterprise data being leveraged includes a complete history of all candidates selected and hired, their key attributes, how they were on-boarded once hired, and their eventual performance in the organization. An analysis engine extracts key features that contributed to candidates’ success and creates a recommendation engine that can rate new applicants along their likelihood to thrive at the organization.

Simple data analytics, right? Yes, except that the algorithms, rather than people, decide which factors matter and which do not. Furthermore, the system continually processes ongoing results of those candidates, updating its recommendation engine rules over time. The system learns from actual experience, just like humans do. But it does so far more rapidly and objectively.

“Now, extend this capability to other high-value, high-frequency business processes,” Hollis writes. “Timing and pricing of supply chain purchasing. Negotiating discounts on large orders. Measuring the temperature of your customers to determine when a small issue might become a big one. Today’s AI-informed recommendations become tomorrow’s advanced automation.