Attracting and retaining AI talent is challenging everywhere. China is no exception.
In our 2022 global AI survey of senior business executives, 75 percent of respondents in China reported challenges in hiring data scientists.1 More than half also said they had difficulties finding talent to fill other critical AI-related roles, including data engineers, data architects, and machine learning engineers, all of whom are needed to design, build, and industrialize advanced digital and AI capabilities.
Our latest research suggests it’s about to get harder, even with recent market contractions. By 2030, demand in China for individuals skilled in building AI products will grow sixfold from current levels (from one million people to six million) as companies race to unlock the more than $1 trillion in potential value AI can deliver in China.2 (See sidebar, “About the research,” for details on our research methods.)
However, local and overseas universities and existing top-tier talent are estimated to supply only about two million (or one-third) of the AI talent needed by 2030, resulting in a shortage of four million people for AI roles (Exhibit 1). Beyond 2030, our research suggests that the declining birth rate will further tighten the availability of AI talent, as fewer students will be entering university programs.
With this enormous talent gap looming, we set out to understand what companies could do to ensure they have the people and capabilities they need to compete. Our surveys and interviews with more than 100 top companies in China reveal two crucial insights:
- Not all talent gaps are the same. While all companies will need to upskill existing employees and look beyond traditional hiring practices to attain the talent and capabilities they need, the types of investments and interventions required will differ based on each company’s level of digital maturity.
- While Chinese graduates increasingly favor local companies and their innovative and performance-based incentive structures, foreign multinational companies operating in China will be able to draw from a much larger talent pool if they effectively leverage their global footprint.
In this article, we share the types of talent companies should prioritize at each stage of digital maturity and how they can best access the skills and capabilities they need as AI talent becomes even scarcer.
Talent and skills needs vary based on digital maturity
The digital and AI initiatives driving the most significant value creation across China will require a wide range of advanced technical skills. These skills fall broadly into seven areas: customer experience, cloud, automation, platforms and products, data management, DevOps (a methodology for streamlining software development), and cybersecurity and privacy. While leaders will ultimately need to build out their talent benches in all areas, our research suggests that their priorities will vary based on which common digital-maturity archetype a company fits into: traditionalist, hybrid, or digitalist (Exhibit 2).
Exhibit 2
Digital maturity will dictate talent priorities for companies in China.
| Archetype | Definition | Priority AI talent battleground | Associated skills |
|---|---|---|---|
| Traditionalist | Companies that have just started their digital transformations and typically have small in-house teams | Data management | Data architecture, data engineering, data analysis, and analytics-translator skills |
| Hybrid | Sector incumbents that have invested heavily in digital transformation and have built strong in-house tech capabilities | DevOps | Agile product management, continuous integration/continuous delivery (CI/CD), and microservices |
| Customer experience | Predictive analytics, design thinking, and automated testing and prototyping | ||
| Cloud | Kubernetes, Docker, and multicloud architectures | ||
| Digitalist | Hyperscalers, digital natives, or AI/tech start-ups with strong tech capabilities | Cybersecurity and privacy | Shift-left security (moving security testing earlier in product development), zero-trust security, and data protection laws and practices |
| Automation | Generative AI, robotic process technologies, machine learning, AI-enabled analytics, and quantum computing | ||
| Unique workforce considerations for multinational corporations | All workers | Fluency in a second language other than Mandarin to leverage and share work across a global footprint Understanding of Western ways of working to communicate seamlessly with colleagues worldwide | |
| Leaders | Strong partnership skills to align efforts with an organization’s global IT and AI standards while effectively executing the expectations of local business leaders | ||
| Product owners | Knowledge of what data and designs can be reused and scaled from different regions, and what must be rebuilt locally to address the uniqueness of the digital ecosystem in China |
Traditionalists
Traditionalist companies are those at the start of their digital transformations. They typically have small in-house teams and face immense competitive pressures to get their digital and AI transformation launched. Much of their emphasis will be on building data foundations to optimize business processes and focusing on targeted use cases designed to deliver immediate business impact (as opposed to research and development of AI capabilities for future innovation). To do so, two types of talent, in particular, are priorities:
- The first type is data management experts with skills in data architecture, data engineering, data analysis, and analytics translation, who can build the data platforms, pipelines, and processes to democratize data access, enable real-time data-driven insights, ensure data quality and governance, and manage the life cycle of use cases. Leaders may hire these experts to serve on data product or use-case teams that bring new digital and AI capabilities to life. Some of these experts will also be needed within a data center of excellence to collectively design and oversee data management processes that ensure appropriate access controls, data quality, and approval and retention policies.
One agricultural enterprise began by building a centralized enterprise data center that supports data management protocols and governance processes and is accessed by thousands of employees from different functions for AI and analytics use cases. The work has significantly reduced IT costs by eliminating the need for repeated development of new data pipelines and has enabled the company to modernize its business methods—for example, by leveraging robots to track breeding conditions for animals and to generate automatic alerts when potential diseases and other issues are detected. - The second type of priority talent is platform and product experts with software development skills who can customize software-as-a-service (SaaS) or other outsourced solutions to improve business efficiency and deliver new customer-facing services. One consumer electronics manufacturer, for instance, began investing in a team of data platform developers after seeing promising results from an initial AI use case, which optimized production schedules and labor productivity. The team will update the underlying models, user interface, data pipelines, and back-end infrastructure so they can continuously enhance and introduce current and additional AI use cases.
Would you like to learn more about QuantumBlack, AI by McKinsey?
Hybrids
Hybrids are those sector incumbents that have already invested mightily in digital transformations. They’ve built strong in-house tech capabilities and foundations and now have set their sights on streamlining development processes to deliver new digital and AI products faster and increase domain expertise to deliver excellent customer experiences. To do this, they’ll need DevOps experts skilled in software development practices such as agile product management, continuous integration/continuous delivery (CI/CD) practices, and microservices that can accelerate deployment. They’ll also compete for customer experience experts who are skilled in using a broad range of predictive analytics, design thinking, and automated testing and prototyping capabilities to launch new customer experiences.
IT efficiency and server spend will also loom large for hybrids as they continue to scale their capabilities and host more AI models and applications in the cloud. Our 2022 cloud survey found that more than three-quarters of companies in China plan to use multiple cloud services, and 90 percent plan to use a mixture of public and private cloud services by 2025.3 Knowing what capabilities are needed and how to run them in each type of cloud will require cloud experts with experience in Kubernetes, Docker, and multicloud architectures.
Digitalists
Digitalists are the digital natives, such as hyperscalers and AI and tech start-ups. While they already have a strong talent bench in most areas of digital and AI, they will need to expand them further to keep up with evolving industry expectations and technology advances.
One area of focus will be cybersecurity and data privacy. In China, there have been increased efforts to mitigate security and privacy concerns that can sideline AI and digital efforts, and digitalists will need experts with a holistic view and a systemic approach to address these issues. At the top of the list are individuals skilled in moving security testing earlier in product development (often called shift-left security), zero-trust security frameworks, and data protection laws and practices.
Another priority will be automation experts with advanced skills in generative AI, robotic process technologies, machine learning, AI-enabled analytics, and quantum computing, who can automate development, testing, and deployment processes end to end to increase the efficiency and speed at which they can bring new capabilities to market.
Multinational corporations in all archetypes
Across all levels of digital maturity, multinational corporations operating in China must also ensure their AI talent has the skills to leverage and share work across their global footprint. For instance, teams will need fluency in a second language other than Mandarin and an understanding of Western ways of working to communicate seamlessly with colleagues worldwide. Leaders will need strong partnership skills to align efforts with their organization’s global IT and AI standards while effectively meeting the expectations of local business leaders. Product owners will require knowledge of what data and designs from different regions they can reuse and scale, and what they’ll need to rebuild locally to address the uniqueness of the digital ecosystem in China.
Consider a worldwide transportation application, developed by a multinational’s European offices, that uses consumer traffic data from Google, Facebook, and Instagram to optimize routes. While most of the company's divisions worldwide could reuse the capability as is, product owners in China would need to guide their teams in modifying the application to pull data from Chinese platforms before they could deploy it.
Addressing gaps by upskilling and diversifying talent sources
Our interviews with companies about their practices to attract, develop, and retain talent suggest that traditionalists and hybrids both have much work to do across most stages of talent management (Exhibit 3). Unsurprisingly, digitalists need to shore up only a few areas to remain on top of their talent management game.
But as we delved further into companies’ strategies, it became clear that the best opportunities for all companies to advance their digital and AI agendas rest in two key areas: upskilling existing talent and diversifying talent sources. In both areas, our research suggests company archetype informs which levers executives should pull.
Upskilling existing workers
Upskilling is a common strategy for sourcing in-demand skills. Our research suggests that companies in China can obtain in-demand skills through targeted capability building in their existing business and AI talent benches (Exhibit 4).
Instead of broad-based efforts, however, our interviews with leaders suggest that upskilling is best targeted to linchpin skills that are hard to find and can’t be outsourced or easily acquired, such as an understanding of the legacy application landscape or existing product functionality (see Exhibit 5 for how to start).
Exhibit 5
To kick off the skills development journey, business executives can consider a checklist of priorities.
| Key actions | Potential approach |
|---|---|
| Prioritize worker skills. Identify skills gaps and devote management time and budget to closing them |
|
| Expand skills training. Provide training needed to cover all workers, preferably with tailored content |
|
| Ensure incentives to train are in place. Introduce training “opt out” system and link with performance evaluation system |
|
| Adjust training content. Offer a mix of “forum, field, and feedback” to improve the effectiveness of learning |
|
| Track impact. Ensure that effective evaluation systems are in place to track effectiveness of training and value for money |
|
| Develop partnerships. Explore partnerships with educators to offer competitive, up-to-date programs and content |
|
| Integrate training into government affairs efforts. Ensure that training provision is integral to government relations |
|
Source: McKinsey Global Institute analysis
The next frontier for AI in China could add $600 billion to its economy
Traditionalists
One such area for traditionalists is crucial analytics translator skills, without which, our work suggests, it’s much harder to gain business buy-in and adoption of new digital and AI initiatives. Upskilling business experts within different domains to identify and evaluate potential digital and AI use cases, assess their potential business value, and guide deployment can position traditionalists to gain value more quickly from their digital and AI investments. This training is best delivered in-house through an analytics academy, an approach that allows organizations to customize content and incorporate active apprenticeships so that business experts can practice what they learn in the classroom.
An advanced manufacturer at the start of its transformation journey, for instance, launched an academy to upskill more than 200 people across the company to become analytics translators. The curriculum included half-day weekly lectures over two to three months on problem solving, talent, and use case requirements; best practices in agile delivery and change management; and on-the-job training guiding use cases from the company’s road map. Since taking on their new roles, these translators have enabled more than 50 new digital and AI use cases.
Hybrids
Only 8 percent of today’s AI talent in China is equipped with advanced AI-related skills, such as edge computing, big data and machine learning, and cognitive AI.4 For hybrids, upskilling existing staff with such skills will be crucial to completing their transformation journeys. But they will need to increase their investment in online courses and certification programs. In our 2022 global AI survey, only about a third of respondents in China reported using such programs (31 percent for self-directed online courses and 29 percent for certification programs).5
One leading financial institution prioritized online learning by providing customized learning journeys based on employees’ roles and career paths. Each employee can access skill-building courses identified as critical to their role via a mobile learning app that offers a broad range of courses—everything from coding in Python to building multicloud architectures to developing the leadership skills necessary for digital transformations.
Digitalists
Digitalists will find their greatest challenge is staying up-to-date with fast-moving emerging technologies, such as generative AI and quantum computing. Encouraging employees to keep up with the latest breakthroughs—for example, by sending staff to academic conferences and giving them time to conduct research, pursue patents, and attend hackathons—can help close their emerging talent gaps.
One technology company gives employees the time, space, and budget to research and develop new capabilities with emerging technologies outside of value-generating use cases. This freedom has resulted in myriad patents and patent applications across AI, blockchain and cloud computing, and new-product innovations.
Diversifying talent sources
Outsourcing work and acquiring essential technology capabilities (and talent to support it) can also provide companies in China with a path toward filling their talent gaps. Multinationals have a distinct advantage in this area given their global reach, which enables them to tap into existing solutions built by colleagues in other regions or new capabilities developed in other countries, such as Vietnam and India. While there are many financial and regulatory considerations in evaluating such paths, such as ensuring adherence to all Chinese data protection laws, our research suggests that some approaches are better than others for each archetype.
Traditionalists
Traditionalists must move quickly to catch up with AI and digital leaders and remain competitive. However, hiring and onboarding new talent to launch their digital transformations, especially in a tight labor market, can take a great deal of time. One way to get AI talent and capabilities quickly is by partnering with vertical IT and SaaS providers. Some leaders use these partnerships to get work under way while they search for new talent. The consumer electronics manufacturer discussed earlier, for instance, outsourced the development of new AI optimization models as it began to build out its talent strategy. By doing so, leaders were able to get new capabilities into production (and generating value) in fewer than eight weeks rather than the many months it can take to onboard new talent.
Others may seek external providers to build their digital system’s entire foundation. For example, a Chinese industrial-vehicles provider engaged a leading software company to integrate more than half a dozen business and factory systems, including those for enterprise resource planning, manufacturing execution, product life-cycle management, vendor management, human resources, and business intelligence. Completed over three years, this work enabled the company to launch a broad set of use cases, including a collaborative product design system that increases R&D efficiency and helps bring new products to market.
When outsourcing work, companies should ensure they have a holistic data and tech strategy aligned with strategic priorities that vendors can use to guide design decisions. In this way, organizations can engage multiple vendors to work on different tasks and projects with confidence that all solutions can seamlessly share data and insights.
Hybrids
Outsourcing can be extremely valuable for hybrids in improving the reach and productivity of their existing tech experts during the next phase of their digital transformations. Outsourcing can free tech staff from the substantial time and effort required to maintain and upgrade back- and middle-office legacy systems.
Today, the rise of enterprise software solutions in China for human resources, finance, communications, and business process automation allows companies to quickly move these systems to the cloud and redeploy AI talent to high-value use cases. In other instances, companies may leverage third-party resources to assist teams in building some part of a new digital or AI solution.
Digitalists
Many digital natives have found frequent expansion and reorganization have led to high tech talent turnover rates and expensive recruitment costs, threatening their continued growth. As the talent gap grows, digitalists will be better served by strategic acquisitions that offer access to new markets or business areas than by building new capabilities in-house. Bytedance’s acquisition of Music.ly is an example. Through this acquisition and subsequent merging of the technology into TikTok, Bytedance gained access to new virtual-reality capabilities that it could use to expand TikTok and to a team of virtual-reality experts capable of using the technology to build out new capabilities.
As demand for AI talent in China increasingly outpaces supply, leaders will need to get creative to ensure they have the people and capabilities to remain competitive in the coming decade. Those organizations that prioritize upskilling of existing talent and strategically leverage outsourcing and acquisitions will find they can close their talent gaps and build their competitive advantage in any market.
