When AI flattens the organization, every layer has to remake the meaning of its role

AI Flattening series directory | hub / Start here (this piece)

hub (this piece) | executives | middle managers | front line | frontline staff

Generative AI is quietly setting off an organizational shift: company hierarchies are being “pulled flat.” From technology giants to traditional industries, AI, multi-agent systems, and data automation are moving into every business process, and redefining the role of every grade. Companies such as Amazon, Moderna, and McKinsey have already begun to thin out middle management, merge departments, and deploy AI agents to automate routine work. The result is the arrival of the age of AI Flattening: junior roles are gradually disappearing, management layers keep shrinking, and the traditional boundaries of responsibility between layers are growing more blurred. Research shows that the average manager now oversees three times as many people as in 2015, and Gartner forecasts that by 2026, 20% of companies will use AI to cut more than half of their management roles. Under this trend, even the top of the house is seeing new AI leadership posts, and a long-standing pattern of power is being recast. This piece looks, in the manner of a business weekly, at five layers of the organization — top management, directors and VPs, middle management, first-line managers, and operational staff — and at how, in this wave of generative AI and multi-agent systems, the meaning of each role is changing, what challenges and openings each layer faces, and what a survival strategy and a path for skill transition look like.

I | Top management

As AI accelerates the flattening of the organization, the role of the senior executive is being repositioned. Imagine an experienced chief executive leading a team that spends months setting a five-year strategy — and an AI system now producing a comparable analysis in six minutes. This is not a science-fiction scene. It is a reality already being played out in boardrooms around the world. The traditional advantages that once dominated the corporate leadership layer (deep industry knowledge, analytical decision-making) are being overtaken by AI’s computing power and insight. Large volumes of data analysis, market forecasting, and other heavy work can be finished quickly by AI. CEOs and senior executives instead have to ask: with AI’s almost “all-capable” assistance, what is the core value of their being there?

The value of top management is shifting from “decision-maker” to “setter of the vision” and “shaper of the culture.” As AI takes on the analytical work that executives used to do themselves in large volume, business leaders need to concentrate more on skills that are uniquely human: an intuition for people and markets, the ability to inspire a shared vision, moral judgment in a messy situation, and the inventive thinking that can improvise. Soft strengths that may once have been treated as accessories when a strategy was being set have, because of the rise of AI, become the deciding point. AI can already do a large share of the work in finance and strategy. What is truly irreplaceable, and scarce, is the executive’s empathy, ethical judgment, and the ability to decide under uncertainty — human capabilities that have long been undervalued. The corporate leader of the future will look more like a “chief philosophy officer,” leading the organization to think about fundamental questions in the age of AI: “who we are, and where we are going.”

In this current of flattening, senior executives face both challenges and openings they have not had before. On one side, decision rights are being pushed down closer to the layer that executes, because AI makes information more transparent, and real-time data can reach frontline teams directly, reducing the need to report up, layer by layer. That means the distance between executives and staff is shortened by AI; the top of the house is no longer high above, but more like a node in a flat network. Executives have to adapt to this change of role, and learn to watch frontline movement directly through an AI platform, to hear staff feedback, and even to take part in the collaboration themselves. On the other side, AI also gives senior executives powerful decision support: leaders can use generative models to simulate market changes and forecast risk, and decisions that used to rest on experience now have data as corroboration. Still, the machine provides answers; people decide the direction. Executives have to choose, among the options AI provides, what fits the company’s values and long-term mission — especially when ethics or corporate social responsibility are involved, a domain AI cannot replace.

The survival strategy for senior executives is to take this change on, rather than retreat behind the curtain. First, the top of the house has to build AI literacy and digital decision-making capability, try the generative tools themselves, and understand their strengths and limits, so that investment and deployment at the strategic level can be right. Many leaders admit that their executive teams lack confidence and understanding of AI (a sign that leadership in the age of AI still has to be raised), and should therefore set an example, put time into learning, and even consider bringing in new posts such as a chief AI officer, so that the decision layer itself has a professional AI voice. Second, senior executives should focus on leadership that is centered on people. When the structure moves from a pyramid to a network, the top of the house is no longer only the one who issues orders, but the one who coordinates the network and enables the team. They need to build a culture of open innovation, encourage staff to use AI to improve the work, and at the same time emphasize AI ethics and risk management, so that the organization does not fall into the trap of blind optimism or misuse. Finally, leaders themselves need a growth mindset, and a willingness to keep adjusting the strategy. Agentic AI is developing very quickly; only if top decision-makers stay agile and take the change on can they steady the company through it. In short, top management in the age of AI will be recast from a symbol of power into a symbol of judgment and responsibility — using AI to drive the business, while guarding the company’s human core.

II | Directors and VPs

For directors, vice presidents, and other upper-middle managers, the flattening that AI brings is both a sorting-out and a chance to recast and upgrade. As the middle layer is thinned, they often have to manage a wider span of business and people directly, and at the same time make sure AI actually lands inside the department. Once the layers of the structure have been reduced, some department heads begin to hold more than one brief: at Moderna, for example, the previously separate resource and technology departments were merged and placed under a chief human resources and digital technology officer, a new kind of cross-domain senior role. In the same way, senior managers in many traditional companies are finding that the map of their duties is being enlarged by AI — they not only lead their original specialist team, but also have to take on digital-transformation work, and even a share of the tasks that used to sit in a parallel department. Upper-middle managers are becoming the organization’s “integration hub”: connecting different functions, combining human and AI resources, and making sure the whole moves toward the same objective.

AI-driven flattening puts new challenges in front of upper-middle managers. The first is the widening span of decision and supervision. When the feedback and the buffer of the middle layer are reduced, directors and VPs may face situation reports from more frontline teams directly. They need to use AI tools well to grasp operating metrics in time — real-time dashboards, automatic analytical reports, and the like — to fill the vacuum left when they can no longer supervise layer by layer. At the same time, they have to be careful not to be drowned by a mass of data, and must build the ability to filter for what is critical. They also have to handle a new problem of cross-department collaboration. AI often leads a company to break the old departmental walls: a marketing team, for example, may use a data-analysis AI for market insight directly, without having to depend entirely on a traditional research department. That requires a more macroscopic view from senior managers, coordinating different teams in the shared use of AI resources, and preventing duplicate investment or islands of resource. And when AI is widely used across departments, upper-middle managers have to design new processes and a governance frame, to make sure AI decisions are consistent and compliant — a new management function that may never have been in their job description.

The challenges are not small, but AI also gives upper-middle managers openings they have not had before. Through AI, they can test hypotheses and drive innovation at a speed that was not available earlier. In the past, a product VP might have needed a team to trial-and-error for months before identifying a market trend; now, by using generative AI to analyze user feedback and competitor movement, insight can be obtained in a matter of weeks. That makes more immediate strategic adjustment possible, and lets upper-middle managers lead the team to turn earlier. They can also put more time into high-value work: once AI has taken on the routine reporting and administrative chores, directors can concentrate on strategic planning, key customer relationships, and the development of the team’s leadership. Especially in the age of AI, the management of people’s emotions and the building of organizational culture have become more important — work that often used not to be a KPI, but is now as important as the technical implementation, because whether an AI change finally succeeds depends on whether people keep up. Upper-middle managers can play the role of a change agent: leading AI-introduction training inside the department, setting up successful examples, and so encouraging the whole organization to take on the new tools. Those who are willing to carry an experimental spirit will create a large demonstration effect for the company, and will also raise their own strategic value in the organization.

Faced with the remaking of the role that AI brings, the survival strategy for upper-middle managers is to recast themselves on purpose. First, they should quickly raise their capability in data analysis and in the application of AI, and at least become familiar with the AI solutions relevant to their domain. That may mean learning again: an operations director learning how to use a supply-chain forecasting algorithm to lower inventory cost, or a marketing director mastering a generative copy tool to improve advertising content. The share of workers with AI skills has already risen significantly across industries worldwide; upper-middle managers should therefore lead in raising their own AI literacy, if they are to lead the team. Second, they need to adjust their style of leadership, from the traditional command from above to authorization and enablement. AI gives a team more autonomy; managers should encourage members to decide, to try, and to innovate, and should themselves step back to provide guidance and resources. Upper-middle managers should act as “coach-type leaders,” with the weight on developing talent and promoting cross-department collaboration, rather than only watching KPI figures. Third, they have to strengthen change management and communication. Because an AI change often comes with organizational adjustment and personnel movement, managers have to talk with the team in a transparent and empathetic way, set out the vision of the change, and ease fear. That not only holds morale; it is also the key to securing resources from above and gathering people below. As some have put it: “Once the organization is flattened, everyone has to learn to take responsibility for a shared task without a clear instruction.” It is exactly this atmosphere of accountability and collaboration that upper-middle managers have to build. In short, the upper-middle managers who dare to take AI on, and recast themselves on purpose, will become a pillar the company cannot do without in the age of AI Flattening.

III | Middle management

If AI Flattening is an earthquake in the structure of the organization, middle managers are undoubtedly at the epicenter. For a long time they have been the hinge between above and below, coordinating the team, supervising execution, and reporting upward. But as generative AI and automation systems gradually replace these functions of coordination and communication, middle managers are facing a full interrogation of the value of their being there. Forecasts show that by 2026, 20% of organizations will use AI to cut more than half of their middle-management roles. In fact, the process has already begun quietly: the number of middle-management vacancies many companies post has fallen significantly in recent years. Some large technology companies have also cut a large number of small-team supervisor posts. Clearly, the middle-management layer is becoming a target of thinning under an AI efficiency-first logic. Many of the “information broker” roles they used to hold have already been replaced by automation tools: AI analyzes data in real time and generates reports, sharply reducing the need to report up, layer by layer; project-management platforms and intelligent agents can automatically track progress and assign tasks, weakening the traditional middle manager’s function of control. When frontline staff, with AI’s assistance, can finish for themselves tasks that used to need a manager’s help, the positioning of the middle manager is bound to be shaken.

Middle management is not, however, walking toward extinction. It is walking toward a rebirth and a remaking. Generative AI will not completely replace middle managers, but it will significantly reduce their numbers and their day-to-day administrative work, freeing them from tedious administration to put themselves into tasks of more value. In other words, middle managers will shift from “the one who issues orders” to “collaborator” and “coach.” As the span of control widens and autonomy at the base rises, the middle managers who remain will instead take on more critical, human duties. They will no longer be the people who look at progress reports every day, because those reports will be generated automatically by AI. In their place, they will act as mentors to the team, attending to people’s growth and psychological state, and putting more energy into skill training and performance feedback. Especially after the organization has cut a layer, frontline staff need human care and a sense of direction even more — if every manager is simply cut, the organization may lose talent because it has lost management with empathy. That is why many companies have not abandoned middle cadres entirely, but are trying to redefine their role: letting them become talent managers and change agents in the age of AI. Some leaders have also stressed that “with artificial intelligence, we can concentrate on collaboration and creation” — what is implied is that managers should put more time into building the team’s creativity and cohesion, rather than being busy with the tedious approval processes of the past.

For the individual middle manager, adapting to this upheaval is full of pressure, and it also holds a turning point. On one side, they bear a sense of career crisis they have not had before: some companies have stopped hiring new managers, have demoted some of the middle layer to individual contributors (ICs), and have asked every remaining manager to oversee more people directly. For these managers, the organizational skills they once relied on to show their value (tight communication up and down, monitoring of the process) seem to have become irrelevant, replaced by an AI system that monitors the metrics and assigns the tasks at any moment. Many middle managers feel adrift: will I be replaced by AI? Is there still a path of promotion? What is more difficult is that they often face a double bind of a rising workload and sharply rising psychological pressure — because even as headcount falls, performance targets only rise, and the organization expects the managers who remain to “do more, and err less” through AI.

On the other side, as in the two faces of the word “crisis”: danger, and opening. If middle managers can recast themselves in time, their experience and skill still have a great deal to offer in the age of AI. First, their deep understanding of how the organization runs, combined with AI tools, can become a multiplier of efficiency. After reducing the management layer, many companies have reassigned middle cadres as specialist advisers or project leads, so that they continue to create value for the team with their technical or business expertise. That shows that a middle manager can, on the strength of industry knowledge and managerial experience, become a source of judgment for the team even without the “title” — especially when AI systems still need a person to calibrate and improve them, and experienced specialist talent is unusually scarce. Second, middle managers can take on the role of an “AI interface” on purpose, acting as a translator between the business and the technology. The landing of many AI schemes needs to be adapted deep in the business situation; if middle cadres can learn some data-science knowledge, they can turn a need on the floor into something the engineering team can use, and the other way around, and so build a bridge for AI application inside the company. Finally, middle managers should turn their focus toward people. However capable AI is, it cannot replace a person in motivating a team or resolving a conflict. Middle managers can use this opening to become shapers of organizational culture: between the strategy set at the top and the execution at the base, they tell the story of the vision, set an example that lifts morale, and build a healthy ecology of collaboration between AI and people. In other words, the person who can understand both the technology and the human heart is the one who can stand out in the new order.

The transformation of middle managers needs the company’s support and their own resolve. For the organization, it should recognize that middle cadres are not wholly surplus; they need to be re-enabled. Companies with foresight will give them a chance to retrain around AI, and will make the path of transition clear — for example, developing some managers into new roles such as agile coaches or data analysts. At the same time, the culture of the company should be adjusted: encourage managers to put down the airs of rank, to be willing to learn new technology, and even to allow a move from a management post onto a specialist track without treating it as a demotion. For the middle managers themselves, what is needed is the courage to renew themselves. Taking on AI tools on purpose is the first step — for example, learning to use a decision-support system for budget planning, or using automation software to simplify an approval process — and feeling, in practice, the benefit AI brings. Second, they have to reposition their own value: a little less control does not mean a loss of value. On the contrary, energy can be put into higher-order management work such as coaching people’s growth and clearing obstacles for the team. When the machine is responsible for “managing the work,” the middle manager has a free hand to “manage the people” and “manage the change.” Finally, it is worth adjusting the mindset, and taking on a wider set of career possibilities. The path of promotion in the future may no longer be sitting, step by step, in a higher management post, but becoming a master in a specialist domain, or becoming a cross-department project leader. In the process of AI remaking the organization, companies need all kinds of talent, not only “managers” in the traditional sense. The middle managers who can see this early and prepare for it have a chance, in the reshuffle, to keep a place, and even to set out again under a new identity.

IV | First-line managers

At the base of the organizational pyramid, first-line managers are the shift leaders, group leaders, or floor managers who lead frontline staff directly. For them, AI Flattening is like a double-edged sword in the hand: on one side, AI tools have lightened many of the tedious day-to-day burdens of management; on the other, they are required to upgrade their skills quickly, so that they can do a new kind of team leadership. As the middle layer shrinks, many first-line managers find that the person they report to has changed from a department manager to a director or VP further up; the span of control has widened, and at the same time the people available to support them have become fewer. If a first-line manager used to supervise ten people, they may now have to lead fifteen or more — and some of the “members” are in fact AI systems: digital “subordinates” such as an automatic rostering system or a performance-monitoring dashboard. That means first-line managers are shifting from a role of only managing people to the managers of a mixed human–machine team.

The application of AI in frontline teams has given the everyday look of a manager’s work a new face. In the past, they had to spend time totaling rosters, watching attendance, and assembling departmental performance by hand. Now, rostering software automatically optimizes the allocation of people, attendance is recorded by a face-recognition system, and the daily performance report is generated by AI in real time. Once many routine acts of management have been automated, first-line managers instead have to face more real-time decisions. For example, when AI detects an anomaly in efficiency at some point on a production line, it may notify the first-line manager at once, and even offer a recommended plan. The manager then has to judge quickly whether the AI’s suggestion is workable, and give an adjustment order. Or an intelligent assistant in a contact center may be handling several customer inquiries at once, and escalate to the floor manager as soon as a question appears that AI cannot answer. In these situations, AI has become an assistant, and it is also pushing first-line managers to keep raising professional judgment and the ability to respond on the spot. In fact, the more widespread AI becomes, the more first-line managers need to play the role of “exception manager”: ordinarily, let AI automate the standard process; but as soon as an anomaly, a human–machine conflict, or the particularity of a single case appears, the manager must step in in time and solve the problem. They therefore need a deeper insight into the department’s work, so that they can quickly find where AI has a gap and take action.

More important still, the age of AI makes the human value of first-line leadership stand out more clearly. Because frontline staff are directly exposed to the shock of AI and automation, they will inevitably feel anxiety, and even resistance. First-line managers, as the leaders who spend their days with the staff, have a duty they cannot hand off: to take on the work of communication and coaching. They need to spend more time in one-to-one conversation with members of the group, to hear difficulties and needs, and to give support and advice. And yet, as some commentary has put it, it is now harder than before to book a deep conversation, because a first-line manager’s calendar may be packed with project meetings and urgent tasks. How to care for people efficiently has therefore become a new problem for first-line managers. They have begun to use AI tools to realize part of the original intent of a “mentoring system” — for example, using an AI platform to collect anonymous feedback and understand the emotional movement of the team, or using a chatbot to answer newcomers’ common questions, so that precious face-to-face time is reserved for the harder issues. Even so, no technology, however advanced, can replace the warmth of a human leader in care. The value of a first-line manager in the team shows more in the details AI is not good at: an encouraging look, a timely word of praise, a fair mediation of a conflict, all of which may have a lasting effect on morale. It is fair to say that once AI has taken over the process, first-line managers have to concentrate even more on the human factor, and become the guardians of the team’s morale.

To grow in this transition, first-line managers need to adjust their mindset on purpose and upgrade their skills.

First, take AI on, and use it well. Learn and use management AI tools on purpose, and hand the tedious work to automation, so that there is time left to attend to the team. That includes a basic literacy with data: being able to understand what an AI report is saying, rather than only receiving it passively. It also means learning how to manage AI “staff”: setting the right algorithmic objectives, supervising the quality of AI output, and stepping in to correct when necessary. A first-line manager may, for example, need to check the conversation logs of a customer-service bot on a regular basis, to make sure the replies fit company policy; or supervise the parameters of AI on a production line, to prevent drift. The team leader of the future is likely to be dispatching several human subordinates and a group of AI agents at the same time — assigning tasks to AI as a commander would, and then letting staff use AI to produce the result. To do that, the ability to orchestrate AI has to be built early, and AI treated as a virtual member of the team.

Second, strengthen communication and empathy, and be the team’s steadying presence. Social and emotional skill does not fall in the age of AI; it rises, because a machine cannot provide the human texture. First-line managers interact with frontline staff every day, and need even more to show understanding and care. That includes talking candidly with the team about the meaning and the necessity of the company’s introducing AI, and quieting rumor and anxiety; and it includes giving patient guidance when someone hits a bottleneck, rather than simply telling them to go and ask the AI. Transparent, empathetic communication is the manager’s last recourse, and the staff member’s only lifeline. Especially after the organization has been flattened, promotion paths shrink and the content of the work changes faster; first-line managers have to be good at helping their people see their own growth — promotion may have slowed, but learning new AI skills can still widen the value of a career. In this, first-line managers should act as shapers of psychological safety, encouraging the team to try AI tools, and not to be blamed even if they fail, and so building an atmosphere of continuing learning. Research has also noted that, when the top of the house clearly supports AI, the share of staff who hold a positive attitude toward AI can rise substantially. The attitude of leadership, then, has a deep effect on the front line. First-line managers should therefore lead the team to take on the AI change with an active, open attitude, and show by what they do: we are learning and improving together.

Third, re-examine the way performance is managed. When AI is deeply woven into everyday work, the traditional way of measuring performance by output has to be adjusted. If, for example, AI-assisted customer-service staff double the number of tickets each person handles in a day, a manager can no longer take “volume handled” as the only standard of evaluation; they also have to attend to service quality and to the degree of AI’s involvement. Performance management will put more weight on the results of human–AI collaboration, rather than on an isolated individual contribution. First-line managers should set new KPIs with the team, so that they reflect the actual look of work in the age of AI — for example, bringing “the number of times an AI tool was used effectively to solve a problem” into the review, to encourage staff to apply the new technology on purpose. In the feedback mechanism, managers can also use AI to analyze each person’s strengths and short boards, and offer more objective advice on development. In the future, first-line managers may no longer be “performance supervisors” in the traditional sense, but more like “discoverers of capability” — using AI to find a person’s potential, and coaching growth according to the person. That not only raises the fighting strength of the team; it also shows the manager’s own irreplaceable value.

In short, AI is freeing first-line managers from tedious administration, and at the same time pushing them toward a leadership post with a higher bar. They need to become all-rounders who understand both the technology and the human heart, and to lead the team to grow in the new normal of human–AI collaboration. Those first-line managers who are willing to take the change on, and to learn quickly, will find that they have not been replaced by AI. Because they can steer it, they will become the most valuable leadership force on the company’s front line.

V | Operational staff

On the front line of the organization, operational staff are feeling the sharp change that AI brings directly. For many people, AI is a new colleague at work, and also a potential competitor: it can help them finish a task, and it may also replace some of the skills they used to take pride in. As companies introduce generative AI and automated processes, a large volume of repetitive, highly specified frontline work is shrinking, and traditional junior roles are quietly disappearing. In one healthcare company with an AI mindset, for example, a team that originally had ten engineers was, after AI was brought in, reduced to a group of three people whose job was to supervise AI agents. For frontline staff, that sends a clear signal: in the future workplace, the pure human demand for “people who only execute” will fall sharply, and will be replaced by people with a compound capability who can collaborate with AI. Several studies have also noted that the skill structure of the work that will be needed will be very different from today’s. This large change is being accelerated in large part by AI, and it also means that every frontline staff member has to keep learning new skills if they are to keep up with what the work requires.

The most immediate change AI brings to frontline staff is a renewal of the way the work is done. Tasks that used to take hours — data entry, writing a presentation, making a chart — may, with the help of AI tools, have a first draft in a few minutes. A customer-service representative can rely on real-time translation and a knowledge base to reply to a customer quickly; a marketing specialist can use generative AI to produce a first draft of a poster or a line of copy, and sharply shorten the time spent incubating an idea. For the individual, the leap in productivity is real: used well, AI tools can raise the efficiency of knowledge workers substantially. Many frontline staff are being freed from tedious, repetitive work, and can put their energy into more creative tasks — talking with customers, designing a new scheme, improving a process. It is fair to say that AI is turning frontline staff from “execution machines” into “creative people” — the machine handles the chores, the human mind concentrates on creativity and human interaction. With AI’s assistance, frontline staff have a chance to take part in tasks that used to sit outside their brief, and to build a more diverse set of skills. That means the meaning of the individual role is being revalued: a marketing specialist may at the same time play part of the role of a data analyst; a sales assistant may step into the early work of a decision recommendation. For those who are willing to take the challenge on, this offers an excellent space for self-development.

The convenience AI brings, however, also leaves many frontline staff feeling adrift and at risk. First is the worry about a living: when they see colleagues around them cut because a process has been automated, who would not be shaken? Especially for those whose skills are narrow, and who do simple, repetitive work, the threat of AI is real. Many companies have used AI to reduce supporting roles such as assistants, personnel, and junior administration. Even without losing the job, some staff feel that their capability has been devalued: they may once have been proud of the speed of their hands or the quality of their memory; now AI outpaces human processing speed in a second. Or the rules of experience they had accumulated are learned by AI in one pass, and never forgotten. This sense of loss — “what I was gifted at has been replaced” — easily leads to a fall in morale. Frontline staff also face the pressure of learning. AI tools appear one after another, and companies often ask staff to take training and adapt to a new system while still doing the work. For some older staff, or those who are not at ease with digital technology, this is like learning a new language, and resistance is natural. After the organization has been flattened, the path of promotion for frontline staff also becomes blurred. The career ladder that used to be clear (for example, becoming a group leader after a certain number of years, and then climbing further) may no longer exist, and is replaced by lateral development: continually widening the range of skill in the present post, or turning toward a new functional domain. For people used to the traditional model of promotion, adapting to this change of idea also takes time.

To keep a competitive position in an AI-driven workplace, frontline staff have no other way than continuing to learn and to recast themselves. The key strategies have several sides. The first is to upgrade the skill mix. Every staff member should build AI literacy, including being able to use the AI tools relevant to their work, understanding basic methods of data analysis, and even learning a little simple program adjustment. The global workforce is rapidly raising the AI-skill content of its work; if you stand still, you will be overtaken by those who come later. At the same time, it is even more important to strengthen the human skills AI cannot replace — interpersonal skills such as communication and coordination, teamwork, negotiation and persuasion, which remain essential to a company’s success. So while frontline staff learn to use the new technology, they should not forget to build soft strengths such as expression and cross-department collaboration — these will be the distinction from AI. A customer-service representative who can show high emotional intelligence, and settle a customer’s feelings in a difficult complaint that AI cannot handle, is showing a value that cannot be replaced. In the same way, a factory technician who, besides operating a machine, can effectively lead a small group to improve a process and lift morale is important in a way no AI robot arm can match.

Second, frontline staff should take the change on, on purpose, and be participants in the organization’s change rather than bystanders. When the company introduces a new AI system, it is worth applying to try it, or taking part in a related project, and raising your own capability in the actual work. On the improvement of a process, or the adaptation of an AI tool, frontline staff often have the most standing to speak — they are close to the actual operation, and can offer valuable views. Daring to speak, and contributing an idea, can turn you from “the one being remade” into “a co-designer.” It not only increases a sense of achievement; it also lets management see that you are willing to learn and grow. Research has also noted that when staff cannot get an AI tool, many will go and find an alternative themselves and use it. This self-starting spirit is exactly what companies need. Frontline people can therefore treat themselves as active problem-solvers: when a need appears that AI cannot cover, they can look, on purpose, for another technical or human path to solve it, and show an ability to adapt. Over the long run, the people who take AI and change on will have a stronger sense of security, and more resilience at work, than those who wait passively for their fate.

Finally, frontline staff should use the organization’s resources well to seek growth. In a period of AI transition, many companies will offer resources such as on-the-job training, online courses, and internal rotation to help staff recast themselves. Do not miss these openings; take part in the training on purpose, and apply what is learned to the work. At the same time, rather than treating the thinning of the management layer as a barrier to promotion, treat it as an opening for development: since the stairs going up are fewer, walk in the wider direction, and build a T-shaped or Π-shaped range of capability across more than one domain. Career development is no longer a single line of climbing, but a networked expansion. Frontline staff can collaborate on purpose with colleagues in different departments, learn something new, and become all-rounders across functions. Of course, the effort of the individual still has to be matched by the organization’s strategy. Some leading companies, for example, have begun to set a “dual-track promotion” path for frontline staff: a technical track and a management track, so that people whose specialist skill stands out can still be promoted and see their pay grow without having to become a manager. Frontline staff should watch for, and use, mechanisms of this kind, and find the development track that fits them best. In a future workplace highly permeated by AI, success is no longer only a contest of titles. It is a contest of combinations of capability. Only by continuing to learn, and adapting with flexibility, can one hold a place in this contest.

Closing and an appeal

As generative AI, multi-agent systems, and data automation move fully into the company, we are living through an organizational change of a kind we have not had before. The corporate pyramid that used to be clearly layered is being worked by the hand of technology into a wrinkled network: from the top of the house to the front line, every role is being redefined. Senior executives are no longer only setters of strategy, but also leaders of direction and guardians of culture. Upper-middle managers fuse more than one function, and become integrators across domains. Middle managers have left the tedious function of control, and have recast themselves as talent coaches who enable the team. First-line managers have turned from executing a process to conducting human–AI collaboration on the floor. Frontline staff are no longer screws on a production line, but knowledge workers who work with AI. Every layer has met both a challenge and a new life, and has forced us to look again at the meaning of “work” and the source of “value.”

In this age of AI Flattening, the only thing that does not change is change itself. If companies and individuals are to grow in this setting, they have no other way than to take the change on. For a company’s decision-makers, it should be recognized that flattening is not merely the cutting of a management layer. It is a thorough remaking of the way the organization runs. That includes putting resources into building staff’s AI literacy, redesigning incentives to reward collaboration and innovation, and building a matching governance frame to manage the risk of AI. Part of the cost saved by thinning management should be put into transition training for the talent already there, to help them upgrade their skills and be equal to a new role. At the same time, leaders have to drive a cultural transition by setting an example: encouraging flatter, more open communication, giving frontline people more decision rights, and building an atmosphere in which people at every layer work together for the mission.

For professionals at large, this change is unsettling, and it is also full of promise. Some traditional posts are disappearing, but new openings are appearing — specialists in the AI domain, analysts of data-driven decisions, managers who can integrate the strengths of people and machines, will all be in high demand in the future workplace. Each of us should become a lifelong learner, and update the structure of our knowledge on purpose. While we learn to use AI tools, we should go still more deeply into the human domains a machine cannot replace: creativity, empathy, the ability to solve complex problems. The organization of the future will lean more toward a network structure, and will need staff who can collaborate “laterally” with flexibility, rather than only climbing upward. In other words, success will no longer depend on how high you have climbed in the hierarchy, but on whether you can cross a traditional boundary and connect more than one skill and resource. The real point is not whether the organization still has managers, but whether each person is willing to take responsibility for a shared task, rather than waiting passively for an instruction.

More deep-dive pieces

The judgment, restated

My judgment is that AI Flattening is not the cutting of a layer. It is a rewrite of the meaning of every layer — first see the hub clearly, then choose the layer you are in and read it in depth.

Three things this week

  • Against this piece’s map of the five layers, mark the layer in your organization where the “tremor is strongest.”
  • Find one process that is still “passed up the chain, layer by layer,” and turn it into a decision on the spot plus a replay.
  • Call one cross-layer alignment: the executive boundary, middle-manager orchestration, the frontline path, and the frontline-staff template.

Next step: start with the middle-managers piece — the epicenter of flattening often falls in the middle, where the pressure is greatest and so is the opportunity. You are welcome to subscribe to Future Insights Lab, and have the next piece in the series arrive in your inbox.

訂閱 Future Insights Lab

不定期寄送 AI × Business 短篇筆記,協助你掌握實務重點。隨時可取消訂閱。