People in the Self-Driving Company

Since ChatGPT brought language models into everyday working life a little over a year ago, we hear the same question in almost every conversation with executives: if software takes over more and more tasks, what remains for the people in the company? I already answered this question in my book "The Self-Driving Company" (Springer 2023; German original "Das selbstfahrende Unternehmen", Springer Gabler 2021) with six theses on people in the company of 2035. This article explains the theses, shows how roles change from leadership to the shop floor, and derives from that what executives can do today.

One year of ChatGPT and the question of what remains

The debate about artificial intelligence has shifted this year. For a long time, the question was whether software could take over demanding tasks at all. Now that language models produce text, analyses, and program code in usable quality, the question is: what should it take over, and what should it not? Policymakers have responded as well. On December 8, 2023, the European Parliament and the Council reached a political agreement on the AI Act, the first comprehensive legal framework for artificial intelligence in the European Union. Among other things, the draft classifies systems for recruitment, promotion, termination, and performance evaluation as high-risk applications for which human oversight is to be mandatory.

For the self-driving company, this is not a contradiction but a confirmation. The basic position of the book is this: software algorithms take over the steering and executing tasks; the companies of the future will continue to be run for and by people. Two of the book's seven central theses set the frame: people will continue to work in and for companies and fulfill the empathetic and creative tasks, and companies will still serve people in 2035. How algorithms replace classic processes is something I described in the article The End of Processes, Long Live the Algorithms. This article is about the people who work in this company.

The vision initially triggers fears in many people. I understand that. Anyone who checks orders, compiles reports, or processes applications every day hears "automation" first of all as a threat to their own job. The history of technological progress, however, shows a recurring pattern: the railway and the telegraph initially destroyed jobs and, after years and decades, produced new professions, new industries, and better working conditions. I expect a similar shift in the years up to 2035: software works through the routines, produces the reports, and runs the simulations; people make sure that it receives the right inputs, pursues the desired goals, and complies with ethical and ecological standards along the way.

Six theses on people in the company of 2035

From an in-depth examination of the vision, six theses on the role of people crystallized in the book. They describe not what is technically possible but what people want, and that changes considerably more slowly than technology.

1. People want to buy products from people. The personal sales conversation will not be replaced in 2035 either; it will be prepared by software.
2. People want to be advised by people. Instead of phone queues and forms, the self-driving company will be able to afford people in expert functions. The trend is already visible, for example in well-founded training programs for product experts.
3. People want to buy goods made by people. Alongside automated production, new personal services and new crafts emerge.
4. People want to work with people. Interaction between human and machine, symbolized by the Excel spreadsheet, decreases; collaboration between people increases.
5. People want to spend time with people. New models for working hours and workplaces create the space for this.
6. People will always have jobs. New professions emerge, and many forms of socially valuable work that have received little recognition so far, such as parenting, caregiving, nursing, and volunteering, are upgraded.

The companies of the future will continue to be run for and by humans like us. Software algorithms will only ever perform tasks that direct and implement.

The six theses share one consequence: what software does better belongs to software; what people want from one another stays with people. Repetitive and highly analytical tasks are already performed by computers faster, with fewer errors, and on the basis of considerably larger amounts of data. They do not tire, do not fall ill, and do not leave for a competitor. It would be uneconomical to continue having people perform such tasks, and it would also be inhumane, because nobody finds fulfillment in transferring data from one system to another.

Why handcrafted work becomes affordable again

The third thesis raises an economic question: if goods are produced automatically and therefore more cheaply than ever, who can still afford craftsmanship and personal services? Today it is usually cheaper to purchase a machine than to employ a person, because human labor is heavily taxed. The problem and its solution therefore lie primarily in the political framework. In the book, I take the position that the taxation of human labor must be abolished if people are to be able to afford services from other people. Whether this happens by shifting the tax burden to manufactured goods or in some other way is a political decision; I consider the direction inevitable.

The result would be a double effect. Automatically produced everyday goods become cheaper for broad parts of society, which lowers the cost of living. At the same time, a growing market emerges for consulting, coaching, care, massage, and "handmade." For a hand-made bread roll, consumers will still be willing to pay significantly more in 2035.

That automation makes things not only cheaper but better can already be observed today. In the fast-food sector, automated production lowers unit costs to the point where higher-quality ingredients become possible at the same consumer price. Robotics provides inexpensively manufactured vacuum cleaners and lawn mowers that precisely map even angled areas. At the same time, the complexity of products increases, and this added value compensates for part of the falling unit costs: companies gain contribution margin and productivity, customers gain variety and benefit. How this value creation comes about is something I described in the articles Value Creation in the Self-Driving Company, Part 1 and Part 2.

How roles in the company change

The theses remain abstract unless they are broken down to concrete roles. The book describes five roles and how their everyday work shifts by 2035. The pattern is the same everywhere: routines and controls disappear; relationships, creativity, and judgment gain weight.

I would like to highlight three of these shifts.

Leadership gains time for what matters. A good managing director ideally divides the day into three parts: one third for employees, one third for customers, one third for owners and peers. Reality is often email-driven hectic activity and constant checking of whether instructions have been carried out. In the self-driving company, operational and tactical decisions are automated and simulated in their medium-term effects. What remains is a clear head for strategy, vision, and the conversations with the people who matter to the company. How these decisions come about is something I described in the article How Self-Driving Companies Decide.

Middle management becomes a coach. When self-learning algorithms make the small everyday decisions, a large share of today's management positions becomes obsolete in its current form. The activity profile of the remaining managers is thereby upgraded in quality: they attune teams to new tasks, listen actively, facilitate, and find creative solutions that go beyond the capabilities of the software. Self-organizing teams do not need control; they need motivation and leadership. The structure required for this is something I described in the article The Organizational Structure for True Agility.

Work adapts to the phase of life. The self-driving company recognizes recruitment needs automatically from its own data, for example when someone leaves the company or a major order comes in, and derives precise profiles from them. Matching, training planning, and promotions take into account not only performance data but also wishes and needs. The working-time model follows life: high capacity shortly after training, less around the birth of a child or toward the end of working life, a year off for a trip around the world without organizational drama. At the end of a task stands a qualification certificate that opens the next step; termination becomes part of a new beginning. What counts is not workplace and attendance, but results.

This is precisely where the human oversight that the draft AI Act provides for personnel decisions comes in. The software prepares recruitment, development, and promotion; the decision about people stays with people. For me, this is not a regulatory burden but the fourth thesis on its way into legal form: people want to work with people, and that includes having a human being sitting across from them when their future is at stake.

What executives can derive from this today

Twelve years lie between the present and the company of 2035. Anyone who takes the theses seriously can do four things today:

• Make routines visible. Determine what share of working time in your areas goes to transferring, checking, and reporting. In our experience from projects, executives significantly underestimate this share; the survey is the first step toward relief.
• Invest in human capabilities. Empathy, conversational culture, facilitation, and creativity can be learned but not bought. Those who start today will have the managers the self-driving company needs in 2035.
• Align working models with life phases. Results orientation instead of attendance, flexible working hours, and a free choice of workplace are not concessions but the prerequisite for the right people to come and stay.
• Plan for human oversight from the start. Every AI system that affects people needs a defined point at which a human decides, and transparency about how the recommendation came about. The AI Act will demand this; good leadership demands it already today.

Personally, one point is especially important to me: the vision of the self-driving company is not a vision of a world without people, but of a world in which people are freed from technically pointless and boring work and are deployed according to their passion and abilities. In working with many companies, I have seen how satisfaction rises when people spend their time with people instead of with forms. More satisfaction creates more motivation, and motivation remains the key to success in 2035 as well.

The next step

The six theses, the five roles, and the employee life cycle are described in detail in the chapter "Humans and the Self-Driving Organization" of my book: To the book. Since November 2023, it has been available in English under the title "The Self-Driving Company" (Springer 2023). A summary of the model can be found in the first article of this series, The Self-Driving Company: A Summary of the Book. If you would like to know which routines in your company can be handed over to software first and which roles change as a result, book an expert meeting.