In 2023 2024 was the year the world discovered generative AI (gen AI), but 2024 is the year organizations will begin to fully embrace and derive business value from this new technology. In the latest McKinsey Global Survey on AI, 65% of respondents reported that their organizations use gen AI on a regular basis, nearly double the percentage from the previous survey just 10 months ago. Respondents' expectations for the impact of gen AI are just as high as last year, with three-quarters predicting that gen AI will bring significant or disruptive change to their industry in the next few years.
Organizations are already seeing real benefits from using Gen AI, with business units that have adopted the technology reporting cost savings and revenue surges. The study also provides insights into the types of risks Gen AI poses — most notably inaccuracy — and the new practices top performers are using to mitigate those challenges and capture value.
AI Adoption Surge
Interest in generative AI is also drawing attention to broader AI capabilities. For the past six years, adoption of AI by respondents' organizations has hovered around 50%. In this year's survey, adoption has jumped to 72% (Figure 1). And that interest is truly global: In the 2023 survey, AI adoption didn't reach 66%. Any Region; however, this year more than two-thirds of respondents in almost every region every Respondents in the region say their organizations are using AI, and by industry, professional services is seeing the largest increase in adoption.
Responses also suggest that companies are leveraging AI in more parts of their business: Half of respondents said their organizations are adopting AI in two or more business functions, up from less than a third of respondents in 2023 (Figure 2).
The adoption of Gen AI is most common in functions where it can create the most value.
Currently, the majority of respondents report that their organization, and themselves, are using Gen AI. 65% of respondents say their organization uses Gen AI regularly in at least one business function, up from one-third last year. The average organization using Gen AI has it most heavily in two functions: marketing and sales, and product and service development, as well as IT (Figure 3), the two functions that previous surveys have determined could generate the most value from Gen AI adoption. The largest increase from 2023 is in marketing and sales, with reported adoption more than doubling. However, across functions, only two use cases in both marketing and sales are reported by more than 15% of respondents.
Gen AI is also making its way into respondents' personal lives. Compared to 2023, respondents are significantly more likely to use Gen AI at work and even more likely to use Gen AI in both their work and personal lives (Figure 4). The survey found that use of Gen AI is increasing across all regions, with Asia Pacific and Greater China seeing the largest increases. Meanwhile, most senior-level respondents are seeing a significant increase in their use of Gen AI tools at work and outside of work compared to their middle management peers. Looking at specific industries, respondents in the Energy & Materials and Professional Services industries report the greatest increase in their use of Gen AI.
Generational and analytical AI investments are starting to pay off
The latest survey also shows how different industries are allocating their budgets to Generational AI. Responses show that across many industries, organizations are roughly as likely to invest 5% or more of their digital budget in Generational AI as they are to invest in non-generational analytical AI solutions (Figure 5). However, across most industries, a higher percentage of respondents said their organization is spending 20% or more on analytical AI than generational AI. Looking ahead, the majority of respondents (67%) expect their organization to increase their AI investments over the next three years.
Where are these investments paying off? For the first time, our latest survey examined the value created by business function from the use of Generation AI. Human Resources is the function where respondents have seen the most cost savings. Respondents most frequently reported significant revenue increases (5% or more) in supply chain and inventory management (figure 6). For analytical AI, respondents most frequently reported cost savings in service operations (consistent with last year's results) and significant revenue increases from using AI in marketing and sales.
Inaccuracy: The most recognized and experienced risk of using Generation AI
As companies begin to realize the benefits of Generation AI, they are also beginning to recognize the various risks associated with this technology. These risks range from data management risks, such as data privacy, bias, and intellectual property (IP) infringement, to model management risks, which tend to focus on inaccurate outputs and lack of explainability. The third big risk category is security and misuse.
Some organizations are already experiencing negative impacts from using Gen AI, with 44 percent of respondents saying their organization has experienced at least one negative impact (Figure 8). Respondents most frequently reported inaccuracy as a risk that has impacted their organization, followed by cybersecurity and explainability.
Our previous research found that there are several governance elements that can help responsibly scale the use of Gen AI, but few respondents reported having these risk-related practices in place: For example, only 18% said their organization has an enterprise-wide council or board of directors with the authority to make decisions about responsible AI governance, and only a third said Gen AI risk awareness and mitigation controls are required skill sets for their technical talent.
Harnessing the power of next-generation AI
In our latest research, we also sought to understand how and how quickly organizations are adopting these new generation AI tools. We discovered three archetypes for implementing new generation AI solutions: Recipient Use a publicly available, ready-made solution. Shaper Customize these tools with your own data and systems Maker Survey results show that across most industries, organizations find off-the-shelf products that are applicable to their business needs, but many are pursuing opportunities to customize models or develop their own (figure 9). Nearly half of the reported uses of Gen AI within respondents' business functions leverage off-the-shelf public models or tools, with little or no customization. Respondents in Energy & Materials, Technology, and Media & Communications are more likely to report heavily customizing or tweaking public models or developing their own purpose-built models to address their specific business needs.
Most respondents reported it took one to four months from project inception to deploy Gen AI in production, but the time it took varied by business function (Figure 10) and by how those functions were acquired. Not surprisingly, the use of highly customized or proprietary models was 1.5 times more likely to take more than five months to implement than an off-the-shelf public model.
AI Generation Outperforms Despite Challenges
Gen AI is a new technology, and organizations are still in the early stages of exploring its opportunities and scaling it across functions. So it’s not surprising that only a small number of respondents (46 of 876) attributed a significant percentage of their organization’s EBIT to Gen AI adoption. Still, these Gen AI leaders are worth taking a closer look at. After all, they are the pioneers, and already attribute more than 10% of their organization’s EBIT to the use of Gen AI. 42% of these high performers say that more than 20% of their EBIT comes from the use of non-generative, analytical AI, and they span a wide range of industries and geographies, but most of them are organizations with less than $1 billion in annual revenue. The AI-related practices of these organizations can provide guidance for those looking to adopt Gen AI and create value in their organizations.
First, Gen AI high performers use Gen AI in more business functions, an average of three functions compared to an average of two for other organizations. Like other organizations, these organizations are most likely to use Gen AI in marketing, sales, and product or service development, but they are also much more likely than other organizations to use Gen AI solutions in risk, legal, compliance, strategy, corporate finance, supply chain, and inventory management. They are more than three times more likely than other organizations to use Gen AI in activities ranging from accounting document processing and risk assessment to R&D testing, pricing, and promotions. Overall, while roughly half of the Gen AI applications reported within business functions use publicly available models or tools, Gen AI high performers are more likely to implement heavily customized versions of those tools or develop their own underlying models than they are likely to use these off-the-shelf options.
In what other ways are these high performers different? First, they're paying more attention to Gen AI-related risks. Perhaps because they're further along in their journey, they're more likely than other companies to say their organization has experienced every negative impact from Gen AI that we asked about, from cybersecurity and personal privacy to explainability and intellectual property infringement. As such, they're more likely than other companies to say their organization considers those risks, as well as regulatory compliance, environmental impacts, and political stability, as associated with using Gen AI, and they say they're taking steps to mitigate more risks than other companies.
Gen AI high performers are also much more likely to say their organizations follow risk-related best practices (Figure 11). For example, they are nearly twice as likely as other companies to involve legal departments and incorporate risk reviews early in the development of their Gen AI solutions (i.e., “shift left”). They are also much more likely than other companies to adopt a wide range of other best practices, from strategy-related practices to scaling practices.
In addition to experiencing the risks of adopting Gen AI, high performers also encountered other challenges that may serve as red flags for others (Figure 12). 70% said they experienced challenges with data, such as defining processes for data governance, developing capabilities to rapidly integrate data into AI models, and lack of training data, highlighting the critical role data plays in capturing value. High performers are also more likely than other companies to have experienced challenges with their operating model, such as implementing agile ways of working and effective sprint performance management.
About the Research
The online survey was conducted from February 22 to March 5, 2024, and received responses from 1,363 participants representing the full range of regions, industries, company sizes, functional specialties, and tenures. Of the respondents, 981 said their organization has implemented AI in at least one business function, and 878 said their organization uses Gen AI regularly in at least one function. To adjust for differences in response rates, the data was weighted by each respondent's country's contribution to global GDP.
