A few years ago, big-data hype resounded through the tech industry, even as many companies frankly didn’t get it. They had data lakes, but few were fishing anything valuable from them. Then artificial intelligence came along promising to finally drive usable insights. But many remain without a practical playbook; do they throw random AI tools at data and wait for it to go kaboom with insights?

“In my experience, this ‘Oh, I’m going to transform my business in one big boil-the-ocean exercise’ does not work,” said John Thomas (pictured), distinguished engineer and director, analytics, at IBM Corp.

The hype around AI may have inflated people’s sense of what it could accomplish. To be sure, AI is potent and can make a huge difference in how people work and how businesses serve customers and earn revenue, etc. But at this stage, people should take a bite at a time to gauge where in their business AI might be useful, according to Thomas.

“I think we are past the hype cycle, and people are looking at, ‘how do I implement successful use cases?'” Thomas stated.

Thomas spoke with Rebecca Knight (@knightrm) and Paul Gillin (@pgillin), co-hosts of theCUBE, SiliconANGLE Media’s mobile livestreaming studio, during the IBM CDO Summit in Boston, Massachusetts. They discussed practical AI in the enterprise and the idea of AI governance. (* Disclosure below.)

AI governance

Wading into a sea of data and working backward toward a use case is a good way to waste a ton of time. Instead, start with a bounded use case, Thomas suggested.

“Once you have a clear understanding of the use case and success metrics for the use case, do you have the data to support that use case?” he asked.

Companies might possess a portion of the needed data but need to acquire third-party or unstructured data from outside the organization. Once they have the data, they have to ask if an algorithm is properly trained to extract accurate insights. IBM is investing in “AI governance,” which vets AI and machine-learning algorithms for bias.

“We have got actual mechanisms in place that IBM research is developing to look at bias detection,” Thomas said. One is the open-source Python library called AI Fairness 360 —  a set of “fairness metrics” for data sets and models, as well as explanations for metrics and algorithms that mitigate bias.

Watch the complete video interview below, and be sure to check out more of SiliconANGLE’s and theCUBE’s coverage of the IBM CDO Summit. (* Disclosure: TheCUBE is a paid media partner for the IBM CDO Summit. Neither IBM, the event sponsor, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE