AI Drives Enterprise Decision-Making, Despite Functional Challenges

If there’s doubt about whether artificial intelligence has reached a tipping point in business, a recent report attempts to settle that question.

The recent KPMG report, “Living in an AI World,” says that artificial intelligence (AI) has reached critical mass in enterprises. Two-thirds of respondents, for example, believe that AI is moving at an appropriate pace within their organizations. As a result, the report finds that today, AI drives enterprise decision making on the ground. The report highlights these benefits emerging from AI in the enterprise:

• Nearly nine in 10 respondents, in myriad industries, said that they expect AI will help their organizations run more efficiently.
• Eighty-nine percent of health care respondents believe that AI has already helped create efficiencies in the industry, and 80% of retail respondents say the same.
• Eighty-five percent of financial-services respondents say that they believe AI will successfully detect fraud.
• Eighty-two percent of respondents in transportation believe that consumers will drive autonomous vehicles in a decade.

There are “challenges,” the report found, but they are largely functional in nature and involve these areas. One somewhat telling finding is that more than one-half of respondents believe that AI is more hype than reality today.

The report authors believe this suggests there’s still a delta between expectations for AI and real-world experience in implementing the technology.

KPMG’s Sreekar Krishna

“Executives underestimate the time and effort required to derive value from AI,” said Sreekar Krishna, principal, innovation and enterprise at KPMG. “Value doesn’t necessarily begin with completion of a production-scale system. It comes from continuing to run the system, and as your processes are transformed.”

The report also noted that there are several key challenges for AI as it becomes more integrated into enterprise operations:

• How to reskill the workforce to exploit AI. C-level leaders are most confident in the ability to retrain workers (79%), but managers are far less confident (38%) about prospects for reskilling. Organizations also need to take a procedural approach to integrating AI into enterprises.
• The need to address security and privacy issues. Roughly three-quarters of respondents see AI as a potential threat to data privacy and security.
• How to address governance so AI can deliver on its promise. Only one in four (25%) of organizations have a mature governance framework in place. But 90% of organizations believe that enterprises should institute an AI ethics policy to govern potential algorithmic bias, and so on.
• How to overcome what KPMG describes as the “failure to launch” problem underlying AI — where it is still confined to lab experiments, not production-ready deployment.