- Just 17% of companies are using artificial intelligence at scale, according to enterprise AI adoption research from KPMG released Tuesday. The report is based analysis of job postings and media coverage and in-depth interviews with senior executives at 30 Global 500 companies that "buy and deploy AI."
- Scaling the advanced technology is a key objective for executives in coming years, according to the report. Companies expect to increase AI investments, related to talent and technology, 50-100% in the next three years.
- Among the surveyed organizations, five companies had more mature AI capabilities. Each had, on average 375 full-time employees dedicated to AI and spend $75 million on AI talent, according to the report. In the next three years, the companies expect to have between 500 and 600 full-time employees working on AI.
AI will play a major role in determining "winners and losers" in industry, KPMG said. And half of companies are looking to the chief information role to lead overall AI strategy.
What's emerging in AI is the rise of the haves and the have nots, and it's caused by the level of enterprise investment. There is almost a "ten-fold gap" in the amount of resources dedicated to AI between those companies with mature vs. early stage AI capabilities, according to KPMG.
Because the technology is in its early stages, the gap is not as noticeable in industry, but that will change as AI matures and scales.
Spending is at the very top, but most companies aren't doing much with AI investment, Cliff Justice, principal, Innovation & Enterprise Solutions at KPMG, told CIO Dive. While "it's the very early days," people understand the coming impact.
As it stands, there are three main uses for AI, Justice said:
Deliver insight, adding to what data and analytics can do, but in a more probabilistic way.
Augmentation of human capabilities, as seen with with virtual assistants.
Full automation, which will have the most "consequential impact economically."
Industry is a ways from full automation, but advancements in AI implementation hint at the technology possibilities. Success comes in the form of newly derived insights or the time saved from robotic process automation (RPA) implementation.
It's not so easy for industry fixtures. Though leading companies including Unilever and Procter & Gamble have troves of data, legacy technology acts as a hurdle and data is often not in the right form for AI.
Barriers arise from agreements with customers dictating how a vendor can use its data, according to Justice. In other cases, if companies are looking for a strong ROI or a small use case, they have to first transform data from non digital formats.
That's one of the reasons Justice cautions over-hyping expectations. AI systems are not plug and play, and instead require enterprises to rethink their approach. But if companies are too leery of early failures, it could set them back in the long run.
Right now, AI lacks maturity, and "I just don't think companies are ready," Justice said. But the mindset is changing and business are starting to adapt.