The CIO Survival Guide to AI

TwitterLinkedInGoogle+Facebook

Artificial Intelligence is paving the way for CIOs to move from operations to driving transformation.

By Kevin Delaney, Senior Writer, Connected Futures


Artificial intelligence (AI) is driving cars, winning at Jeopardy, and if you follow science fiction, it’s getting into all sorts of trouble on its own.

But CIOs can’t worry about rampaging robots in a far-off future. They need to know what AI can do for their organizations. Today.

“Three years ago, you might have been too early to start on AI,” said Dan Olley, CIO of Elsevier, “now if you wait three years you will be left behind.”

AI (along with its close cousins, machine learning and predictive analytics) promises to streamline IT, while enabling a new dimension in customer insights and experiences.

Experiences that IT will need to innovate.

“I think it’s a huge opportunity,” said Niel Nickolaisen, CIO of O.C. Tanner. “If I want to change how my role is perceived by the organization, if I want to be perceived as a strategic business leader, one of the best ways is to deliver an analytics home run.”

But how do CIOs make the AI revolution work for them? And where do they start?

“The mistake is to hire a bunch of nerds and hope that good things happen,” said Vasant Dhar, a professor of information systems at New York University’s Stern School of Business. “It has to start top down with solid business thinking.”

“The important thing,” he said, “is what is it going to do for the business? Is it going to make better products, help retention, reduce churn? What is it going to do, and why?”


Thinking Machines, Keeping the Lights on

For the IT organization, simply “keeping the lights on” remains as important (and challenging) as ever. AI technologies can play a key role in streamlining clunky processes, freeing IT to concentrate on innovation.

“Agility and complexity cannot coexist,” said Nickolaisen. “We have to clean up a lot of the complexity that we have built over the last 20 years.”

Dhar touts AI’s ability to handle massive amounts of unstructured data, discerning key patterns and insights along the way. The true paradigm shift, he said, is when systems learn on their own, without waiting for humans to write the programs.

“AI can do this,” he said. “The tech has come a long way. The tools are more accessible, and what took a month now takes a day.”

Olley sees many practical applications for AI in IT, especially in automating tedious tasks. “Things like log analysis, server balancing, security threat detection,” he said, “you can see how machine learning can really make an impact.”

Once that traditional domain is automated and simplified, IT can then concentrate on using AI to drive innovation.

“You can move business insights to the next level with these technologies,” Olley offered as an example. “With algorithms that learn and recommend when customers are about to churn.”


Start Early, and Start Tiny

One thing is certain: AI will drive profound changes in the coming years.

To avoid being disrupted, Dhar said, “it’s important to get started early.”

Fully integrating AI technologies into IT, however, can seem daunting. Especially when IT budgets are flat.

To prove the business case for further investment, Nickolaisen said, “you don’t want to start small, you want to start tiny.”

“There are some very cost effective ways to start to play in the sandbox of these advanced analytics tools,” he added.

Nickolaisen recommends taking your data sets to a university or analytics user group where AI is being explored. Even a small amount of prize money, he said, can spur some great work. His team also takes advantage of vendors that encourage first-time users to experiment with their AI tools.

“We’ve done some really compelling analyses,” he said, “that shook the organization dramatically.”

Again, the name of the game is building a clear and compelling case for new investment.

That begins with data.

“You need to have the architecture, the structure that is going to get you that data,” said Olley. “It’s like training a human. You can’t have unsupervised learning with bad data.”


‘Just Enough to Be Dangerous’

Culture change, to drive a constant spirit of inventiveness and curiosity, is also critical.

“Learn just enough to be dangerous,” Nikolaisen joked.

He has encouraged his team to learn more about chat bots, by reading, experimenting and talking with experts. The end goal is to prove that chat bots could be a novel way to interact with customers.

“If it turns out there is a business case,” Nikolaisen said, “we will add chat bots to our product roadmap and go forward.”

Olley stresses that an understanding of the new technologies is important across the organization. “Getting people to spot problems that can now be solved is half the problem,” he said.

“It’s an exciting time to be in IT,” Olley added, “but you need to understand the art of the possible or you will get leapfrogged. We are at that juncture.”