How Leaders Are Helping Employees Integrate AI Into Their Work

Almost all organizations learned the hard way that  generating optimized outcomes from technology required far more than plug and play. 

With AI, the organizational disruption is likely to be more challenging, time-consuming and expensive. Interestingly, the crucial variables do not appear to be AI-specific, but rather, human skill sets that already exist within the organization, such as good communication, strong interpersonal relationships nurtured over time, the ability to discern context - and domain expertise which guides effective implementation. JL 

Nada Sanders and John Wood report in Harvard Business Review:

Leaders have been inundated by AI’s potential to transform work. The future involves humans working with AI. Leaders feel pressure to “do something” to implement AI solutions. But how does their workforce integrate AI to achieve business outcomes? The current hype is around AI capability and organizational change, but it is the human ability to understand context - which AI lacks - that necessitates interpersonal skills, such as conflict resolution and communication (as well as) domain expertise preserving knowledge among experienced talent and developing it among young workers. “Companies acquiring AI without a new business model (are) digitizing a horse and carriage." Competitive advantage cannot be achieved without humans in the loop.

Right now, the corporate world is in the throes of a generative AI hype cycle. Leaders have been inundated with speculation about the technology’s potential to transform work. However, this raises an entirely new sets of problems for leaders: They now understand that the future will involve humans working with AI and feel immense pressure to “do something” to implement AI solutions. But how do they actually integrate AI with their workforce to achieve good business outcomes?

Since the launch of generative AI tools such as ChatGPT in November 2022, and the tsunami of other applications that have followed, we have conducted dozens of in-depth CEO interviews across industry sectors. Our goal was to identify how generative AI has changed the practices of leading firms and extract lessons for success.

Our recent interviews reveal a marked shift in focus from earlier studies — both our own and those of other researchers. While our own earlier interviews showed the importance of a human-focused culture, our recent interviews have signaled a definitive shift toward development of individual human competencies. The current hype cycle is around AI capability, digital skills, and the need for organizational change management, but our current finding reveals the importance of specific human capabilities to the successful use of AI.

In fact, it is the human ability to understand context — which AI tools lack — that necessitates the need for greater human skills. This is an important insight that can help leaders understand the necessary human elements that drive an organization’s successful use of AI.

The Skills Companies Really Need

We observed two categories of human skills that leaders see as critical — and which organizations are actually teaching to their employees. First are effective interpersonal skills, such as basic conflict resolution, communication, skills of disconnecting from emotions, and even mindfulness practices. Second is domain expertise, with a focus on preserving that knowledge among experienced talent and developing it among young inexperienced workers.

 

While digital literacy was viewed as a given by the leaders we spoke to, interpersonal skills such as the ability to effectively communicate, meaningfully engage with others, and garner team cooperation were not. Our interviews reveal that while technical skills get more press, it is these uniquely human skills that companies need — and find to be in short supply in today’s marketplace. Peter Cameron, CEO of Lenox Group, underscored that workplace success depends on the ability to cultivate personal relationships, saying that, “Nothing replaces long-term relationships that are personal — and the longer the relationship the better.” Maria Villablanca, co-founder and CEO of Future Insight Network, pointed to specific qualities to look for in talent: “People that can be creative and innovative in the way they find solutions — problem solvers.” Broader research backs this up: A study of 1,700 global companies found that companies that excelled on human capital metrics were four times as likely to have superior financial performance.

One of the greatest values of experienced workers is domain expertise — deep knowledge of one’s environment. As AI takes over more tasks, there is a significant danger of atrophy of skills and loss of this kind of knowledge. John Sicard, president and CEO of the supply-chain-management company Kinaxis, pointed out that domain knowledge is essential to helping a business navigate through turbulent times. He gave the analogy that when an airplane experiences anything unusual, the autopilot is immediately turned off and pilots take control — at which point the pilots need to know what to do. Developing and maintaining these skills is essential.

Further, generative AI is shown to be more useful as a co-pilot for senior employees that can sift through AI “hallucinations” — inaccurate information presented as fact — and take the output as an aid. Inexperienced employees, however, may not be discerning enough and need a path to develop this knowledge. The sentiment was echoed by Ted English, former CEO of TJX Companies and current executive chairman of Bob’s Discount Furniture, who told us leadership requires “a lot of instinct, experience, and knowledge. Some of it you can’t get from a machine. Technology reinforces and allows you to make a more confident decision.”

While much has been written on the need to combine humans with AI, the real issue is how to implement this in practice. Here’s what companies need to do.

More Than Just Combining Humans and AI

In 2020 we published our Four I framework for using AI that calls for creating a human-centric organization. As a result of our current findings, we have modified the descriptors of our original framework, making it more specific in its mandates. This offers a direct comparison — and shift — between the use of AI in the post pandemic business environment.

The first layer of the framework is intentionality. In this context, we mean that a company’s business model should be purposefully designed around AI capability, rather just applying AI to existing processes. Spencer Fung, president and CEO of Li & Fung, a global supply-chain and logistics company, gave us an analogy: “Companies acquiring AI without a new business model is like a company digitizing a horse and carriage — while the competition has created a digital automobile.” He offered the clothing company Shein as an exemplar: a company that became a fast-fashion industry leader by using a business model centered around a seamless e-commerce platform leveraging a data-driven approach.

Next comes integration across all functions of the enterprise, with horizontal communication and AI as the enabling layer — in other words, getting rid of silos. For example, ekaterra, the world’s largest tea company, invested in a unifying AI platform to create complete horizontal connectivity, integration, and real-time visibility.

The real challenge, however, is implementation. Garry Kasparov, former world chess champion, has written that winning performance does not come from combining the best technology with the best people — but from the best process of combining. To achieve this, talent must be familiar with AI capabilities and know how best to utilize them.

However, AI is an evolving technology, and that necessitates a business add slack to the system to allow opportunity for learning. One of the authors learned this in an earlier project with Lockheed Martin on the F-22 Raptor, where downtime was needed for workers to become familiar with new tools and technologies. With an evolving technology new learning may even cause a temporary loss of productivity, but that loss will ultimately be more than made up — it’s the proverbial, “putting down an axe to learn how to use a chain saw.”

Implementation also requires differentiation between “deep work” and “shallow work,” with the latter assigned to AI. Deep work, however, requires worker training for developing talent literacies. For example, one company builds in 90-minute sessions as part of the regular work week.

The last part of the framework is indication, or performance measurement. Forget traditional performance measures or the inane debate on working from office or home. Look to develop novel metrics tied directly to purposeful intentionality of your business model using generative AI. Organizations from Google to Schneider Electric are now using AI to redefine what to measure, how to measure, and improve performance.

Humans With Enhanced Skills “In the Loop”

Competitive advantage cannot be achieved without humans in the loop. Rushing to replace talent with AI is a huge mistake.

Why?

First, AI is copyable. What is not copyable is a unique business model, processes, and thoughtful integration of humans.

Second, AI is based on historical data that may not hold true in a volatile global business environment. Kinaxis’s Sicard underscored that “every math-based model collapsed when the pandemic hit. None of the assumptive parameters could be trusted. It is not an indictment on the science … but an indictment on the believe that those technologies eliminate the need to be agile.” Business decisions are not made in a vacuum separate from issues of labor, inflation, geopolitics — and this is where domain expertise of people is essential. In our executive interviews we repeatedly heard that the new competitive advantage is “the human mind,” “human critical thinking,” and “human decision intelligence.”

Third, AI is subject to hallucination and “drift,” where output is either fabricated by the AI or simply inaccurate. Handing the reins over to AI poses great risks in terms of inaccurate decision making and even legal liability.

AI is still a tool. The centerpiece are people, but with enhanced human literacies, a well-thought-out business model, and superb processes that integrate humans with their AI co-pilots.