Recruitment has more than doubled since 2017, The percentage of organizations using AI has plateaued between 50-60% over the last few years. The array of companies getting the best financial returns from AI continues to lead the competition. The results show that these leaders are making greater investments in AI, engaging in increasingly sophisticated practices known to enable rapid AI development at scale, and AI talent shows signs of doing well in a tight market. In terms of talent, we took a closer look at AI hiring and upskilling for the first time. The data show that there is significant room to improve AI team diversity, and consistent with other research, diverse teams correlate with better performance.
table of contents
- Five Years in Review: AI Adoption, Impact, and Spending
- Mind the gap: AI leaders stay ahead
- Talent Tales in AI: Hot New Roles, Continuing Diversity Issues
- About research
1. Five Years in Review: AI Adoption, Impact, and Spending
This is the fifth year in a row that we have conducted a global survey of the role of AI in business, and we have seen a change during this period.
2. Mind the gap: AI leaders stay ahead
Over the past five years, we’ve been following AI leaders (we call them AI High Performers) to see how they differ. We see more signs that these leaders are expanding their competitive advantage than evidence that others are catching up.
First, we don’t see an increase in leader group size. Over the past three years, we have defined AI high performers as those organizations where a respondent sees the greatest revenue impact from AI adoption, i.e., 20% or more of his EBIT from the use of AI. I’ve been The percentage of respondents who fall into this group has stabilized at around 8%. The survey results show that this group achieves superior results primarily through AI boosting top-line profits. Because they are more likely to report that AI boosts revenue than it reduces costs, even though they also report that AI reduces costs.
Second, high performers are more likely than others to follow core practices that drive value, such as connecting AI strategies to business outcomes (Figure 1). Importantly, they are also more frequently involved in “frontier” practices that enable the development and deployment of AI at scale, or what is called “industrialization of AI.” For example, leaders are more likely to have data architectures that are modularized to enable rapid response to new AI applications. They also often automate most data-related processes. This improves the efficiency of AI development and provides higher quality data to feed into AI algorithms, increasing the number of applications that can be developed. Also, an AI high performer involved a non-technical employee in creating her AI application with a new low-code or no-code program that allows companies to speed up the creation of his AI application. He is 1.6 times more likely than other organizations. Over the past year, high performers have used more standardized tool sets to create production-ready data pipelines or used end-to-end platforms for AI-related data science than other organizations. You are more likely to follow certain advanced scaling practices, such as , data engineering and application development developed in-house.
High performers may also have a head start in managing potential AI-related risks that other organizations have not yet addressed, such as individual privacy, fairness, and equity. Overall, we have seen little change in organizations reporting awareness and mitigation of AI-related risks since the survey began four years ago, but AI high performer respondents They are more likely than organizations to report engaging in practices with known AI-related risks. Helps reduce risk. These include ensuring AI and data governance, standardizing processes and protocols, eliminating manual errors by automating processes such as data quality control, testing model effectiveness and monitoring potential issues. .
Investment is another area that could contribute to widening the gap. High performers in AI are poised to continue spending more on AI efforts than other organizations. Respondents from these leading organizations, like others, say they will invest more in the future, but they are currently spending more than others. An AI High Performer respondent is nearly eight times more likely than her peers to say their organization spends at least 20% of his digital technology budget on her AI-related technology . And these digital budgets make up a much larger percentage of corporate spending. The AI High Performer respondent reports that his organization spends more than 20% of his total enterprise revenue on digital technology, five times more than her other respondents.
Finally, all of this could be pushing AI high performers in attracting AI talent. These organizations have shown that hiring for roles such as AI data scientists and data engineers is not too difficult. Respondents from organizations that are not AI high performers say it is “very difficult” to fill these roles than those who are AI high performers.
Bottom line: High performers are already well positioned for sustained AI success, greater efficiency in new AI development, and consequently a more attractive environment for talent. The good news for organizations outside of Leadership Groups is that there is a clear blueprint of best practices for success.
3. The Tale of AI Talent: Hot New Roles, Continuing Diversity Issues
Our first in-depth study of the AI talent landscape demonstrates the maturity of AI, reveals the most common strategies organizations employ for sourcing and upskilling talent, and explores the link between diversity and success. Again, it sheds light on the issue of AI diversity.
Recruiting is hard, but not so hard for high performers
All organizations report that hiring AI talent, especially data scientists, remains challenging. High performers in AI report slightly lower difficulty, hiring some roles, such as machine learning engineers, more often than others.
Retraining and upskilling are popular alternatives to hiring
When it comes to sourcing AI talent, the most popular strategy among all respondents is reskilling existing employees. Nearly half do. Recruiting not only from top universities, but also from lesser-tier technology companies, such as regional leaders, is a common strategy. However, a look at the strategies of high performers shows that organizations are best served by leveraging as many recruiting channels as possible (Figure 2). These companies are doing more than others to recruit his AI talent from various sources. Survey results show that while they are more likely to hire from top engineering colleges and technology companies, they are also more likely to source talent from other universities, training academies, diversity-focused programs and professional organizations. increase.
Responses suggest that both AI high performers and other organizations are upskilling their technical and non-technical employees on AI, with both AI high performers and other organizations nearly half of respondents said they are reskilling as a way to attract more AI talent. However, High Performer is taking more steps than other organizations to build her AI-related skills in her employees.
High performers are nearly three times more likely than other respondents to say their organization has a capability-building program in place to develop the AI skills of tech workers. The most common approaches they use are experiential learning, voluntary online courses, and certification programs, while other organizations mostly rely on voluntary online courses.
High performers are also much more likely than other organizations to do more than provide access to voluntary online courses to upskill their non-technical employees on AI. . High performers are almost twice as likely as other respondents to report offering peer-to-peer learning and certification programs to non-technical people.
Improving the diversity of AI teams is a work in progress
We also surveyed the level of diversity within organizations’ AI-focused teams and found that most organizations have significant room for improvement. Only 27% of respondents’ organizations have an average percentage of employees who identify as female on these teams (Figure 3). If you look at the average percentage of racial or ethnic minorities developing AI solutions, the share is similar, just 25%. Additionally, 29% of her respondents said their organization does not have minority employees working on her AI solutions.
While some companies are working to improve the diversity of their AI workforce, more is being done to improve gender diversity than ethnic diversity. He said 46% of respondents said they wanted to increase gender diversity within teams developing AI solutions through procedures such as partnering with diversity-focused professional bodies to recruit candidates. states that the organization has an active program of A third of organizations say they have programs to increase racial and ethnic diversity. Additionally, organizations with women and minorities working on AI solutions often have programs in place to address the experiences of these employees.
Like previous McKinsey studies, this study shows a correlation between diversity and outperformance. An organization where a respondent perceives at least 25% of their AI development workforce to be female says she is 3.2 times more likely to be an AI high performer than other organizations. Companies where at least a quarter of their AI development employees are racial or ethnic minorities are more than twice as likely to be high performers in AI.
About research
online survey will be on-site from May 3 to May 27, 2022 and from August 15 to August 17, 2022, spanning the full range of geographies, industries, company sizes, job specialties and tenures. We collected responses from 1,492 participants representing Of those respondents, 744 said their organization employs AI in at least one function he had and were asked about their organization’s use of AI. To adjust for differences in response rates, the data are weighted by each respondent’s country’s share of global GDP.
