"AI has been around for a long time," says Stuart Frankel, CEO and co-founder of Narrative Science. "While it is super-hot and very buzzy today, there are still some success stories of AI that we just don't consider AI anymore. We use it every day and we're used to it. I think that's a natural progression. Once that esoteric technology that's sort of hard to explain gets wide adoption, it's no longer AI anymore."
In its survey of 345 business executives from a variety of industries, fielded by National Business Research Institute from April 25 to May 27 this year, Narrative Science found that 26 percent of respondents said they are currently using AI technologies in the workplace to automate manual, repetitive tasks. That's up 15 percent year-over-year. And 38 percent reported using AI technologies in total. But when Narrative Science asked organizations about their use of technologies that rely on AI techniques — predictive analytics, automated written reporting and communications, voice recognition and response and so on. — 88 percent of respondents who had said their organizations don't use AI reported their organizations use one or more of those technologies.
In other words, companies are benefitting from AI-powered applications without even realizing it.
But what you call it is less important than what you're doing with it, Frankel says.
"I don't think in most cases that it matters, except to say that people that are making technology decisions, decisions that impact big areas of organizations, need to have some understanding of the technology they're buying," Frankel says.
Whatever term organizations use for these technologies, they are using them, according to Narrative Science's study. Thirty-eight percent of the survey group say they are already using AI technologies, and 56 percent plan to deploy AI technologies within the next two years (23 percent within the next 12 months). That means 62 percent of respondents' organizations will likely be using AI technologies by 2018.
Frankel points to the fact that AI technologies have become much more accessible as a big factor in their explosion within the enterprise.
"It's easy for departments now or even individuals within certain organizations to test AI technologies," he says. "They don't need to get IT involved. I think what we're seeing is AI penetrate the enterprise pretty quickly, but not necessarily widescale across many companies."
IBM Watson is a great example, Frankel says.
"Two years ago, three years ago, testing Watson was a multi-million dollar, potentially multi-year initiative for most organizations," he explains. "There are only so many companies that can take on that kind of commitment. Watson has followed a wide-sweeping trend to make web services available, make technologies available via a series of APIs. That makes it much easier to test. If organizations can test it and get value out of it, that leads to much greater adoption downstream."
Of the AI technologies currently in use, predictive analytics dominates. The survey found 58 percent of respondents are using a combination of data mining, statistics, modeling and machine learning to analyze current data and make predictions about the future.
"Certainly one of the key headlines from our research is that predictive analytics is really moving into the enterprise very, very quickly," Frankel says. "Companies are starting to see the real value in their data. They're making decisions against their data instead of just viewing their data."
Narrative Science notes that organizations are likely gravitating toward predictive analytics because of the tremendous potential it offers across many industries, whether it's preventing costly hospital re-admissions in the healthcare sector or reducing unplanned downtime and allowing for more efficient supply chain management in the manufacturing space.
Research firm Gartner predicts 40 percent of the new investment made by enterprises will be in predictive analytics by 2020.
The study found that companies that place a priority on innovation tend to generate the most value from their technology investments. Fifty-four percent of respondents said their organization has an innovation strategy, and 62 percent said their organization has a dedicated innovation budget. Of those organizations with an innovation strategy, 63 percent believe they are skilled at using big data to solve business problems. Only 13 percent of respondents from organizations without an innovation strategy felt the same way.
Of the companies that have an innovation strategy, 61 percent use AI to identify opportunities in data that would otherwise be missed. Only 22 percent of respondents without a strategy said the same.
Unsurprisingly, the survey also found the data science talent shortage remains top of mind: 59 percent of respondents said the shortage of data science talent was a primary barrier to generating value from their data. Of the survey respondents who said they have deployed big data technologies, about roughly 50 percent said their organization is skilled at using big data to solve business problems. Of the group that said they were skilled, 95 percent use AI technologies. That's up from 59 percent last year, which Frankel says is an indication that companies are turning to intelligent systems to augment their data science capabilities in response to the shortage.
"To the extent that the knowledge of data scientists can be codified into a machine, organizations can get a lot of leverage out of it," he says.
In fact, he notes, the idea of a partnership between human and machine extends well beyond data science.
"One of the major compelling differences in this year's survey compared to last year is the changing perception of AI technologies," he says. "This year, the conversations pertaining to AI are about the power of the partnership. There is growing awareness that AI technologies, when combined with human skills, is optimal because it produces results reaching beyond what either group could achieve alone."