We’re living in a new age of AI and automation. Robots and computers, long used for performing routine physical tasks, are now more cognitively capable than ever. Businesses are increasingly adopting these technologies to aid and scale deep thinking and analysis. For instance, nearly two-thirds of 1,000 business executives we surveyed said they plan to increase their investments in AI. We do, too, and we’ve already begun.
Our Technology & Digital function works inside McKinsey to innovate firm tools and processes, ensuring we bring the best to our clients. All of our people in the function are encouraged to solve problems, pursue new ideas, and launch new products.
We felt that these innovations could take more advantage of AI or automation to achieve their aims. After some assessment, we realized that many of our colleagues in the function had an opportunity to learn more about AI and new techniques, such as natural language processing and machine learning. We decided to invest in scaling these skills and capabilities.
The state of AI in 2021
Our approach
Our firm is a recognized industry leader in AI. We began this effort by embarking on a dual journey to add more data scientists at McKinsey, while at the same time upskilling our colleagues who already had some exposure to AI and automation.
Hiring talent is important; new people bring new ideas and they can help quickly fill gaps in capabilities. But upskilling our existing talent is perhaps even more critical. Doing so provides them with the skills they need to succeed as the shift to AI continues, reduces disruption in existing methods, and helps our people feel motivated and engaged. Investing in existing talent also helps with successful deployment. After all, no one understands how to get things done—and make them stick—in your organization better than your own people.
We developed an ambitious plan to upskill more than 500 colleagues in different roles for AI/machine learning techniques and their applications. The roles included leaders, business sponsors, product managers, user experience professionals, engineers, data experts, and support professionals. We also formed a focused team of about 20 AI experts and commenced the journey.
The teams started with a concept of AI pods—mini three to four member teams of AI experts—that supplemented our product teams. Each focused on five elements throughout their assignments: defining and refining our AI strategy; convening allies; incubating and delivering use cases; building capabilities; and the dissemination and radiation of lessons learned to help upskill colleagues across the function.
From our own client experience, we know that a one-size-fits-all training module would not be an optimal approach. Instead, it would have to cater to colleagues with varying degree of needs for AI techniques in their work. This led to the development of three levels of training, each with varying degrees of complexity and depth. They include:
- AI Aware: Every day, new techniques are getting invented to automate and augment processes through AI. To make our colleagues aware of these possibilities and evolving changes, we curated and packaged a few online and in-house courses and materials. This “AI for beginners” package is designed for everyone in the firm to understand AI applications in business and the power of data for AI use cases.
- AI Ready: Once a colleague is “AI aware,” they can spot AI opportunities. But to engage with AI experts or to plan an AI initiative, colleagues need more than just the basics. Dependency on data, data science techniques, and ROI are some of the variables which make it challenging to engage with AI. So we ran custom boot camps led by internal practitioners to demonstrate broader, newer technologies for machine learning and AI. Further, this was supplemented by a series of online trainings to embark on a continuous learning journey.
- AI Capable: This is the toughest skill to develop. With this level, we wanted our tech talent to be able to build complex AI solutions and play roles like AI researcher, product manager, AI user experience expert, machine-learning architect or engineer, data engineer, data analyst, or a junior data scientist. Ahead of the intensive six-month, on-the-job course, we first aligned the curriculum with each participant’s priorities and then ensured that they and their leadership recognized the ROI of the initiative.
What we learned and achieved
Through this program, we saw that successful upskilling happens quickest for people already in relevant product development roles: technology product managers up-skilled to AI product managers; software developers to machine-learning engineers; or data architects to data engineers. We aligned individuals with relevant ongoing projects to have hands-on experience and paired them with AI capable experts they could shadow and learn from.
This combination of dedicated oversight support from our internal “AI Capable” colleagues and help from external experts on nuanced techniques enabled further upskilling. Within six to nine months, we had them capable of an intermediate level of independent execution. An enthusiastic group of software professionals also chose to move to data science through a formal external data science course.
We feel the effort has been hugely successful. In addition to the thousands of client-facing colleagues who work on AI for our clients, we have had about 100 Technology & Digital function colleagues qualify as “AI Ready” through this program. They now possess an intermediate knowledge of the AI product lifecycle, including topics such as data dependency, data infrastructure and architecture, data science techniques, and AI ROI.
Meanwhile, more than 50 colleagues in the function are now “AI Capable,” and possess advanced skills to build complex AI solutions. These colleagues can now play roles like AI product manager, AI user experience expert, ML engineer, and data engineer or data scientist. And finally, more than 500 Technology & Digital function colleagues are now “AI Aware;” they understand AI applications in business, the power of data, and common AI use cases.
These are colleagues who are upskilled for life, who have grown and developed in a way that will sustain them in their careers for the long term. That’s the kind of change we seek to create for our clients every day, and in this case we’ve made it possible by giving our people what we feel they are owed—skills that are fit for the future.
The authors would like to thank McKinsey partner Angelika Reich and digital innovation and design leader Brandt Flomer.