Originally posted on hbr.
Generative AI tools have taken the world by storm. ChatGPT reached 100 million monthly users faster than any internet application in history. The potential benefits of efficiency and productivity gains for knowledge-intensive firms are clear, and companies in industries such as professional services, health care, and finance are investing billions in adopting the technologies.
But the benefits for individual knowledge workers can be less clear. When technology can do many tasks that only humans could do in the past, what does it mean for knowledge workers? Generative AI can and will automate some of the tasks of knowledge workers, but that doesn’t necessarily mean it will replace all of them. Generative AI can also help knowledge workers find more time to do meaningful work, and improve performance and productivity. The difference is in how you use the tools.
In this article, we aim to explain how to do that well. First, to help employees and managers understand ways that generative AI can support knowledge work. And second, to identify steps that managers can take to help employees realize the potential benefits.
What Is Knowledge Work?
Knowledge work primarily involves cognitive processing of information to generate value-added outputs. It differs from manual labor in the materials used and the types of conversion processes involved. Knowledge work is typically dependent on advanced training and specialization in specific domains, gained over time through learning and experience. It includes both structured and unstructured tasks. Structured tasks are those with well-defined and well-understood inputs and outputs, as well as prespecified steps for converting inputs to outputs. Examples include payroll processing or scheduling meetings. Unstructured tasks are those where inputs, conversion procedures, or outputs are mostly ill-defined, underspecified, or unknown a priori. Examples include resolving interpersonal conflict, designing a product, or negotiating a salary.
Very few jobs are purely one or the other. Jobs consist of many tasks, some of which are structured and others which are unstructured. Some tasks are necessary but repetitive. Some are more creative or interesting. Some can be done alone, while others require working with other people. Some are common to everything the worker does, while others happen only for exceptions. As a knowledge worker, your job, then, is to manage this complex set of tasks to achieve their goals.
Computers have traditionally been good at performing structured tasks, but there are many tasks that only humans can do. Generative AI is changing the game, moving the boundaries of what computers can do and shrinking the sphere of tasks that remain as purely human activity. While it can be worrisome to think about generative AI encroaching on knowledge work, we believe that the benefits can far outweigh the costs for most knowledge workers. But realizing the benefits requires taking action now to learn how to leverage generative AI in support of knowledge work.
How Can Generative AI Help?
When it comes to AI, it’s often said that “AI won’t replace you, but a person using AI will.” Instead of automating your job away, the power of generative AI can help to improve your ability to do cognitively-challenging knowledge work.
The key is to use generative AI to manage the flood of information washing past you every day. Humans have limited cognitive information processing capacity. On the other hand, most knowledge workers today are inundated with a high-velocity in-flow of digital information and always-on communications. This crush of information is creating a “digital debt”: an ever-increasing backlog of information waiting to be processed by each knowledge worker.
If you feel this way, you are not alone. Sixty-eight percent of workers in a recent Microsoft survey indicated that they do not have sufficient uninterrupted time to focus on their core activities during the workday. The share of working hours taken up by emails, electronic meetings, text messages, and search for and review of digital content is increasing even more.
This is where Generative AI can help — if you use it wisely. In particular, generative AI can be useful in three major ways: reducing your cognitive load by automating some structured tasks, boosting your cognitive capacity for unstructured tasks, and improving the learning process for your job.
Reducing cognitive load.
Generative AI tools can enhance performance and productivity by freeing mental capacity to focus on hjgher-value unstructured tasks. You can do this by offloading structured and repetitive elements of knowledge work to generative AI tools. In addition to reducing cognitive load, this can also make your job more interesting and satisfying by removing some of the drudgery involved.
Generative AI is already showing benefits in reducing cognitive load across a variety of industries. Attorneys at the global firm Allen & Overy use a system called Harvey to efficiently locate and access case law and draft simple contracts. This allows them more time to analyze complex legal issues and advise their clients.
In marketing and advertising, generative AI can automate routine content generation such as creating product brochures or personalizing email campaigns. A recent BCG survey of chief marketing officers found that two thirds of respondents were investigating generative AI for personalization, and half are exploring it for content generation.
In finance, a corporate bank is applying generative AI to reduce cognitive load from a constant inflow of new externally-generated financial market information. The system quickly analyzes and summarizes annual reports, earning call transcripts, and analyst reports to keep the bank’s relationship managers better informed about important developments. By streamlining the information search and review process, relationship managers have more time to focus and serve their clients.
As these cases demonstrate, delegating some of structured tasks to generative AI can help to relieve the stress of cognitive overload so you can focus on more important tasks. The drudgery can be done faster, and possibly better, by a computer, while you can improve your performance on the tasks that remain.
Boosting cognitive capabilities.
Another approach to augmenting knowledge work is to use generative AI to boost higher-order cognitive processes to perform unstructured tasks. Three important areas are critical thinking, creativity, and knowledge sharing.
For critical thinking, generative AI can help people to ask better questions about the challenges they face. Experimenting in an executive education setting, researchers found that 94% of the time, engagement with AI (including generative AI) led to asking a wider range and variety of questions than the respondents otherwise would. This in turn led to exploring ideas and possible solutions that they may not have considered, likely leading to better performance.
Another study found that ChatGPT was particularly useful in the idea generation and communication phases of strategy process. The AI tool created plausible strategic ideas with high efficiency. Its “storytelling” capability was particularly useful in helping to articulate and communicate the ideas. On the other hand, the tool was less helpful in suggesting ways to implement strategy, possibly due to the tool’s lack of access to detailed information about the company, its capabilities, and other relevant contextual information.
University of Missouri professor Tojin T. Eapen and colleagues detailed how generative AI can promote divergent thinking by making connections among diverse concepts. Generative AI assisted not only in developing new ideas but also in evaluating and refining them based on criteria such as feasibly, impact, cost, and novelty. In a separate survey of more than 1,000 content creators (bloggers, podcasters, and short-form video producers), two-thirds of respondents indicated that they use the tools for creative tasks. Fifty-three percent said using the tools enhanced their creativity and productivity. Furthermore, those who had used the tools had a higher number of followers and had generated higher income.
Beyond generating knowledge, generative AI can also help to share it. Intellectual assets are dispersed across organizations in a wide variety of documents, policies, processes and individual heads, making it difficult for people to access knowledge that already exists in the organization. By leveraging generative AI, companies can bridge the knowledge gap, facilitate knowledge sharing, and empower knowledge workers with know-how they need to excel at their jobs. For example, to assist its wealth management advisors, Morgan Stanley implemented a generative AI model, trained on a vast set of internally captured knowledge and expertise. Making wealth management knowledge readily accessible to every advisor in the company has been “transformative” by empowering advisors to efficiently address their clients’ specific questions and concerns.
Achieving mastery requires practice and not just classroom work. However, to be useful, practice requires feedback on performance — and providing personal feedback for every employee can be prohibitively expensive. The improving capabilities of generative AI are making it possible to consider the idea of an AI mentor for every knowledge worker.
For students, Generative AI can play roles such as advisor, tutor, coach, and simulator. It can provide frequent feedback, personalized instructions and explanations, alternative viewpoints, and opportunities to practice through simulations. For example, Duolingo recently added two new generative AI-powered features to its language-learning software. Roleplay allows users to practice free-flowing conversations in a foreign language, and Explain My Answer that provides constructive personal feedback to users upon request.
Generative AI’s potential role in learning also extends to the work environment. Take, for example, the challenge of becoming a successful call center agent. Gaining skills in this complex, fast-paced environment requires a combination of instruction, practice, feedback, introspection, and immersion. At Fidelity Investments, trainees learn about a topic, then handle calls with a mentor listening and advising, and then meet with the mentor and other trainees to discuss their experience. Generative AI can assist in this process, both during training and after. It can monitor a conversation with a customer and suggest what an agent can say or do to resolve the customer’s issues. It can also answer the agents’ questions later. A recent study of a call center introducing generative AI found that this type of support helped to improve productivity and quality for all workers, while also increasing employee satisfaction. It also accelerated individual learning by enabling novice workers to progress more rapidly along the experience curve compared to those who did not leverage such system.
How Managers Can Help Workers Use Generative AI
Hundreds of millions of people around the world are already using generative AI tools at home and at work. It’s likely that everyone is using generative AI differently, even if they’re using the same tools. This is an opportunity and a threat for companies. Each person’s experiments can generate innovations that could have great value. But some of the experiments may stray into dangerous territory by creating errors, breaking regulations, making biased decisions, or releasing private information to the public.
These tools are entering the sphere of knowledge work, whether wanted or not. Encouraging your people to experiment with the tools and try them in their jobs can help to alleviate some of their worries. This in turn can lead them to take a constructive and proactive stance toward working with AI, not fighting against it. To help your knowledge workers make the most of generative AI, consider these three key actions.
Define policies and assign responsibilities.
While generative AI tools offer promising advantages, they can introduce risks to the organization. Therefore, it’s critical to understand and mitigate the risks of these tools, and ensure that the procedures for safe are followed. For example, using generative AI to write draft emails or summarize documents is easy and potentially highly productive. However, knowledge workers should be told not to do so with private information, as Walmart did earlier this year. Similarly, employees should know that, while these tools are fast and easy, they not always accurate. One does not want to find out about this weakness the way a lawyer did, when a judge cited him for including fictitious case law in his generative AI-drafted legal brief. Also emphasize the biases that are created either in training or applying the tools and suggest ways to alleviate them.
Encourage experimentation and innovation sharing.
Demonstrate the tools’ capabilities in staff meetings and point to interesting use cases inside and outside of your company. As people share their innovations, ask probing questions to help everyone develop an awareness of the risks involved and know how to mitigate those risks. Encourage peer learning, where employees teach each other how to use the tools. Building on that, ask workers to share the innovative practices they’ve developed, and take steps to help others adopt the best practices.
Celebrate the wins.
By encouraging your team to treat Generative AI as a voyage of discovery instead of a defensive activity, you can help them build confidence and capability in using the tool to enhance their cognitive capability and improve their job-specific skills. Make a big deal of the best case uses, and celebrate the innovations and the innovators. By doing so, you can not only improve the way workers use generative AI, but also improve the innovative culture of your group.
Knowledge work has always required ongoing learning to keep up with the progress of innovation and knowledge. Keeping up with the changes wrought by generative AI tools will require not only learning about the tools, but using them to increase your knowledge and hopefully expand your role. In a recent McKinsey report, the nearly 1,700 executive respondents indicated that they expect more employees to be reskilled than laid off.
The time to start working with generative AI is now. Generative AI can be a boon for knowledge work, but only if you use it in the right way. New generative AI-enabled tools are rapidly emerging to assist and transform knowledge work in industries ranging from education and finance to law and medicine. Companies are starting to introduce generative AI-powered innovations into their processes, and to promulgate policies on how to use the tools safely.
However, there is no need to wait for these externally-imposed changes. You can start now to use generative AI for your own benefit, once your understand and learn to mitigate the associated risks. Using free tools already available on the web, you can reduce your cognitive load from constantly rising tides of information, while also boosting your cognitive abilities and learning effectiveness. Now is the time to start using generative AI in your knowledge work, and to help your colleagues to use it wisely.