Companies are using more artificial intelligence and starting to realize some quantifiable benefits. At the same time, the wild proliferation of technology firms claiming to offer AI tools is hindering companies’ efforts to attain game-changing ROI.

That’s one conclusion of a new KPMG study of 751 business decision-makers with at least a moderate knowledge of AI in their industry, across five industries.

In four of the five industries, a majority of participants said they feel that at present AI is still more hype than reality. The most skeptical industry was transportation (69%), followed by retail (64%), technology (57%), and health care (52%). Financial services was the only industry that viewed AI as more reality than hype, with just 42% selecting the latter option.

“There’s an absolute flood of boutiques and startups with the next tool,” Traci Gusher, principal and U.S. lead for innovation and enterprise solutions at KPMG, tells CFO. Often, the tool “has an AI label slapped on it because there’s a machine learning model somewhere in it, but it doesn’t deliver the promised value because it’s not really an AI solution, just a solution packaged as AI.”

The unfortunate trend is “creating disruption in companies’ ability to capitalize on the benefits of AI. There is distrust, which makes it even more difficult to understand what’s a real benefit vs. what might be fake.”

But while the continual flood of new tools certainly won’t abate, Gusher says there’s reason to hope companies will get better at making such distinctions.

For one thing, to date there hasn’t been nearly enough training. Whereas many companies look at AI as simply a technology play, it really should be viewed as a strategic, enterprise-wide initiative. That requires building deep AI capabilities across the organization.

“AI can be a bit like a toddler. It learns fast, but only through education and observation,” says Gusher.

Only 47% of health care companies represented among survey participants, and 52% of retailers, even offer an AI training course. The other industries are doing better — 65% of both technology and transportation respondents, and 57% of those in financial services, offer such a course.

Another emerging trend might also help pave the way for an eventual better understanding of AI.

Guidelines for its ethical use have been issued by organizations like the World Economic Forum and the Organisation for Economic Cooperation and Development. Plus, data privacy measures like Europe’s General Data Protection Regulation and the new California Consumer Privacy Act are creating a more regulated data market, and data is the key ingredient for AI applications.

“We’re going to keep seeing this, and eventually it will move to legislation and regulation of AI,” Gusher says. “At that point we’ll get some clear levers and guardrails saying that if you call something AI, it should exhibit some specific characteristics. It will make it easier for companies to sift through all of the [tools in the market].”

In fact, one of the more surprising findings of the study, considering the level of antipathy directed at government today, was the degree to which respondents favored the regulation of AI.

A majority of participants from all five industries approved of such government involvement. Support ranged from 82% (transportation) to 59% (financial services).

Support for government oversight actually was strongest among those most engaged in and knowledgeable about AI. More than 9 in 10 (91%) of respondents whose companies have made AI fully functional at scale, and 75% of those with “high” AI knowledge, said government should be involved in regulating AI technology

Support for government oversight actually was strongest among those most engaged in and knowledgeable about AI. More than 9 in 10 (91%) of respondents whose companies have made AI fully functional at scale, and 75% of those with “high” AI knowledge, said government should be involved in regulating AI technology.

“Any time you put in any kind of process, there are questions about how it’s going to affect customers, the brand, and profitability and growth,” notes Gusher. “AI is no different. But organizations are struggling to get the talent to do this work at the volumes they want, and they’re nervous whether they’re doing it right. They want some help with that.”

Meanwhile, the study showed a major disconnect between the views of C-level executives and lower-level managers on whether employees are well-prepared with the necessary skill sets to adopt AI. Most (79%) of C-level executives said they believed employees are ready, but only 36% of managers felt the same.

“Executives are far more optimistic, because their CIOs are telling them about all the amazing new technology they’re bringing into the organization,” says Gusher. “Whereas the folks on the ground that are executing processes and responsible for driving value are seeing value but saying, ‘we’ve got a long way to go.’ ”

Indeed, executives underestimate the time and effort required to derive value from AI, according to KPMG. “Value doesn’t necessarily begin with the completion of a production-scale system,” Sreekar Krishna, a principal in the firm’s innovation and enterprise solutions practice, said in the study report. “It comes from continuing to run the system, and as your process are transformed by what the model is doing.”

, , , , , , , ,

One response to “Can Companies Come to Grips With AI Hype?”

  1. I believe it is true that AI , Machine learning and blockchain are now buzzwords, and professional folks like to use them without deeper understanding of these technologies. These technologies do offer solution for many problems but at the same time, we need to have proper planning and educate our workforce so they are ready to take this on. Data has been the center of the business world, and AI and Machine Learning increasingly can take advantage of data and make sense of it through the use of trending tools. In order to take full advantage of the technology, first we need to be honest with ourselves and have a reality check. Get ahead of the curve, train our trainers, and equip them to handle the technology appropriately. It is not a Magic Wand that will solve all of our problems, as it’s “Garbage In, Garbage Out.” If we can’t resolve underlying problems like data silos, no tool can produce the results we all hoped for.

Leave a Reply

Your email address will not be published. Required fields are marked *