hfng - Fotolia

With AI tools, enterprises need to differentiate between hype and value

Enterprises can reap real value by implementing AI applications, but seeing that value through the fog of hype can be difficult, says one expert.

While there may be lots of potential for enterprises that implement AI tools, a lot of the hype that has emerged over the past year may be just that: hype.

"There's still a fairly wide gap between what a lot of the platforms can do and what their advertising says they can do," said Josh Sutton, head of data and artificial intelligence at the advertising consulting firm Publicis.Sapient in Boston.

There's certainly no denying the hype, and there are some good reasons for the excitement around AI tools. Advances in computation power, combined with large stores of training data, have pushed machine learning capabilities forward at a rapid rate over the last couple of years.

But there are questions about how enterprises will translate these computational leaps into operational systems that deliver business value. In Gartner's report, "Hype Cycle for Emerging Technologies, 2016," cognitive systems and machine learning were listed at the peak of inflated expectations. This has led some commentators to speculate that the hype is rapidly outpacing the ability of the technology to deliver.

For Sutton, the problem is in viewing AI as an overarching technology. It's really made up of lots of different domains, such as machine learning, natural language processing, chatbots and predictive analytics. Right now, companies are looking to sell software platforms that combine all of these into one product, but Sutton said software vendors that take a more focused approach and build tools that address a single domain tend to deliver better functionality.

"There's a growing understanding that the terms AI and cognitive are catch-alls for a lot of technologies," he said. "Things are evolving at such a pace where I don't think you can say who the market leader will be."

The real opportunities for using AI tools today, Sutton said, come in the areas of collecting unstructured social media data to generate insights about products and services, developing chatbots for customer service and engagement, and automating knowledge work. The first two uses are utilized today with varied levels of success, and will continue to improve as time goes by.

Automating knowledge work through AI tools will be a bit trickier, but it offers the greatest opportunity, Sutton said. It will allow companies to forge entirely new business models that aren't simply predicated on reducing labor costs, but that invent new ways of tackling business problems, much like how Uber leveraged mobile adoption to make it easier to hail a cab.

"The companies that are using AI to create different models will be the winners going forward because they're creating a competitive advantage," Sutton said. "They're cutting cost out as an added benefit, but [that's] not the reason they're doing it."

Next Steps

Don't expect heavy job losses from growing role of AI tools

AI increasingly pushes into creative job functions

AI chatbots need to be more conversational to engage users

5 prerequisites for applying AI to an ops environment

Dig Deeper on Enterprise applications of AI