Generative AI went mainstream in 2024 as businesses rapidly adopted the technology and began realising substantial benefits, according to a new McKinsey study. The global management consulting firm’s survey of over 1,300 participants found a remarkable 65% of respondents’ organisations are now regularly using generative AI, nearly double the rate from just 10 months prior.
The payoff has been significant cost reductions and revenue increases for companies putting generative AI into use, the study revealed. Human resources was the function most commonly seeing cost decreases over 20%, while supply chain and inventory management had the highest share reporting revenue gains above 5%.
“Organizations are already seeing material benefits from generative AI,” the McKinsey report stated. “The technology’s potential is no longer in question.”
Overall AI adoption also surged to 72% of respondents’ companies, up from hovering around 50% for the past six years as excitement built around generative AI capabilities like natural language processing.
Fueling the adoption frenzy, three-quarters of respondents predicted generative AI will lead to significant or disruptive changes in their industries going forward. Companies are investing heavily, with about as many spending over 5% of digital budgets on generative AI as on traditional machine learning.
While most are still in experimentation mode, a subset of respondents categorized as “generative AI high performers” demonstrate what’s possible with the technology. These 46 organizations attribute over 10% of their EBIT to deploying generative AI.
The high performers tend to use generative AI across more functions, are less reliant on off-the-shelf models, and are much more likely to follow best practices around risk mitigation, data management, and scaling processes.
These are the early movers who realise competitive advantage by customizing models and developing robust operating capabilities around generative AI.
As generative AI saw broader adoption, respondents increasingly recognised risks like inaccurate outputs, intellectual property infringement, and cybersecurity threats. 44% of companies using the technology experienced at least one negative consequence.
“Responsible practices like risk governance and output monitoring must be priorities from day one,” the report warned. Only 18% of respondents have an enterprise-wide council overseeing responsible AI currently.
The study found high performers were more likely to embed legal reviews, bias testing, and risk practices into their generative AI development cycles.
Looking ahead, McKinsey expects the bulk of value creation will require extensive customisation of generative AI models on proprietary data rather than generic off-the-shelf solutions.
“The enterprise of the future needs orchestrated combinations of foundational models – off-the-shelf, open source, and finely tuned to its specific needs, the report added.