Q&A: How to deliver more data value across the enterprise

A s enterprises amass data, they need better ways to organize it to align with business needs. We spoke with expert Neal Fishman about his book, Smarter Data Science: Succeeding with Enterprise-Grade Data and AI Projects, which delves into using information architecture (IA) to make data accessible and valuable for AI and applications in hybrid cloud environments.

Tell us about your book …

Organizations realize that data is an intrinsic asset that needs to be managed. This aligns with the idea that data is the new oil, a new type of energy and asset that can maximize business opportunities.

Increasingly, a lot of peoples’ jobs involve data and data science, so the book is intended for both IT professionals and executives and puts cloud computing and AI into a business perspective. It takes a big-picture look at cloud computing and how organized data is needed for intelligent workflows and the cognitive enterprise.

What about data management?

There’s a perception that data is somehow alive, that it does something. But for me data is inert: it’s not self-aware and it’s not self-organizing. This is why an IA is an imperative to helping AI and other applications make data valuable. IBM Executive Chairman Virginia (Ginni) Rometty said “there can’t be AI without IA.” The book explores this concept.

To make data actionable, some type of host is required. This could be a computer program, a SQL statement, or an algorithm in a machine learning model. In the book, I address moving from being data rich and information poor (DRIP) to effectively organizing data—not just collecting more of it, which doesn’t solve the problem for an effective IA.

You mention making IA adaptive and more flexible …

A common challenge is how IT should build solutions for businesses that are continuously morphing. The book addresses IA in the context of being adaptive so that the architecture can keep up with the changing pace of business.

Consider that a chess master can plan 15-20 moves ahead as well as their opponent’s maneuvers. Similarly, anticipating where IA needs to be to align with your company’s strategy and future planning is key. To help with creating an adaptive IA, the book describes three essential pillars: data flow, data zone, and data topography.

What data science problems do businesses face?

The fact is that data-centric projects are challenging and often fail to meet expectations. Many companies have a data warehouse or a data lake; if one data lake fails, they create a new one. Data lakes typically fall short from an IA standpoint, specifically how to build IA for AI in a prescriptive way.

A specific data science problem is how to add humanity into your cognitive enterprise, which involves improving user experiences. The book delves into how to support a hyper-personalized experience for both the employee and the customer.

With hybrid and multicloud computing, as well as with edge computing, not every piece of data and process needs to happen everywhere; you can optimize processes and workloads to the most appropriate node. We can use architecture to help us with think about how to distribute data for resiliency in complex data topologies.

What does smarter business mean to you?

As we embrace AI, intelligent workflows and the cognitive enterprise in business, it comes down to modeling and organizing data in a manner that helps us meet current and future needs without breaking the assets we’ve already built.

In our modern world, data has the power to dramatically improve lives. In many ways, we are a data-dependent society. Experiencing a lag in access to information or not having the right data can result in dire consequences. Smarter business is about being adaptive and responsive, providing ultra-personalized experiences and being automated and intelligent. It’s recognizing that a dependency on data means treating and managing data as an essential asset.

Source: ibm.com