Every few months, a new headline announces the end of the middle manager. Research by Gartner, cited in Deloitte's analysis, predicts that through 2026, 20 percent of organisations will use AI to flatten their organisational structure, eliminating more than half of current middle management positions. An analysis published in Business Insider and cited by Deloitte found that US employers were advertising 42 percent fewer middle management positions at the end of 2024 than in spring 2022. It is easy to read these numbers and conclude that management itself is on its way out. The reality is more interesting, and more useful for anyone responsible for a team. Having fewer layers does not mean less need for management skill. It means the skills that matter are changing. Companies that work with a corporate trainer in Chennai are noticing a shift in the questions they ask. They are no longer asking whether managers are needed. They are asking what managers should actually be good at. Quick answer: AI is increasingly taking over coordination, reporting, and routine administrative tasks. It is not taking over judgement, coaching, or accountability for quality. Managers now need to know how to check AI-assisted work, make sound calls where data runs out, develop people through changing roles, and redesign work around what AI can and cannot do.
What the Flattening Headlines Get Right and Wrong
The headlines are right that some management layers will thin out. When AI handles status tracking, scheduling, and first-draft reporting, managers may no longer need to spend as much time on coordination. They are wrong to treat a prediction as a settled outcome, and wrong to assume that the work disappears along with the title. Deloitte's 2025 research on the future of the middle manager argues that for most organisations the sensible path is neither eliminating managers nor keeping the role exactly as it was. It calls for reinvention. The same article cites research associated with INSEAD, as reported by Fast Company, which found that 80 percent of transformation programmes led by middle managers succeed, compared with 20 percent of those led by senior management. Managers sit closest to where change actually has to happen.
The Parts of the Job AI Is Taking Over
Managers are not short of tasks that could be handed to software. Deloitte's research found that nearly 40 percent of a manager's time goes on firefighting today's problems or on administrative work. Preparing reports, chasing updates, consolidating figures, and drafting routine communication all fall into that bucket. Delegating this work to AI can be a genuine gain, provided the time saved is put to productive use. That is the part many organisations skip. If the hours saved simply fill with more delivery pressure, managers end up with the same workload and a new set of responsibilities on top.
What Managers Now Need to Know
How to Check AI Output Without Becoming a Bottleneck
AI can produce work that looks polished and still fails to move the task forward. A Harvard Business Review study involving managers at two consulting firms used the term "workslop" to describe this kind of output. The researchers found that managers were expected to catch errors, validate results, and uphold quality standards, usually without any formal training in how to do so. Useful skills here are specific and teachable. They include spotting made-up facts, judging whether a prompt was framed well, and fact-checking AI-generated analysis. A manager who can do this quickly protects quality without reviewing everything line by line.
How to Use Judgement Where the Data Runs Out
Deloitte identifies judgement as the foundation of the reinvented manager role. It describes judgement as the ability to make difficult decisions when no clear answer exists or when information is incomplete, drawing on context, culture, and empathy. AI can summarise options. It cannot take responsibility for the call. This matters because judgement is built through experience, and many newer managers were promoted for individual achievement rather than for decision-making under uncertainty. Practising difficult decisions and receiving feedback, rather than simply reading about them, is how this skill develops.
How to Coach People Through Changing Work
When tasks change, people need more support, not less. Deloitte cites a 2019 survey of 500 workers and leaders across five countries, which found that effectively developing direct reports can raise their performance by as much as 27 percent and make them 1.5 times more likely to exceed their goals. Coaching is one aspect of a manager's role that becomes increasingly valuable as routine work shifts to machines.
How to Redesign Work Around AI
Managers are close enough to the work to see what is happening and far enough away to see how to improve it. A retail fashion example in the Deloitte research makes the point well. Fashion buyers initially resisted AI recommendations because the tool seemed to take away the creative work they enjoyed. Their managers reframed the buyers' role around vision and direction, and resistance faded. That was a human judgement about identity and motivation, not a technical fix.
How to Set Standards for AI Use
Someone has to decide when AI output is good enough, what junior staff should still learn to do by hand, and how to handle questions from clients or customers. The Harvard Business Review researchers found that managers were making these decisions team by team, in isolation, because no firm-wide direction existed. Clear standards are a leadership responsibility, and managers need to be part of setting them.
Old Manager Tasks and the Skills That Replace Them
The table below shows how familiar management activities are shifting, and what each shift asks of the person doing the job.
Compiling status reports - What is changing: AI drafts and summarises reports. - Skill needed now: Interpreting insights and acting on them. Reviewing finished work - What is changing: AI-assisted output arrives faster. - Skill needed now: Verifying quality and spotting plausible errors. Assigning tasks - What is changing: Some tasks move to AI. - Skill needed now: Deciding what people and tools should each do. Annual feedback - What is changing: Roles shift continually. - Skill needed now: Providing regular, personalised coaching. Applying set procedures - What is changing: Routine cases are automated. - Skill needed now: Using judgement to handle exceptions and trade-offs.
The Hidden Strain on Managers
None of this is easy to absorb. The Harvard Business Review researchers describe a pattern they call role elevation, where AI shifts work upward for junior staff and partners. Managers, however, did not experience elevation. Their new responsibilities for oversight, coaching, and quality control were simply added to their existing delivery targets. Without support, managers are not lifted up by AI. They are buried by it. The same study also highlights a perception gap. It draws on BCG survey data showing that executives are roughly twice as likely as individual contributors to describe employees as enthusiastic about AI, which leaves managers caught between leadership ambition and day-to-day reality. Deloitte's survey adds that 36 percent of managers felt they were not sufficiently prepared for the people-manager parts of their role. A caution is worth stating. The Harvard Business Review research draws on 18 interviews at two consulting firms, so it is a qualitative signal rather than a measured trend across industries. It matches what Deloitte's much larger survey describes, though, which makes the direction credible.
Protecting the Leadership Pipeline
There is a longer-term risk too. Junior employees traditionally learned by watching managers structure a plan, test an analysis, and handle a difficult conversation. If managers are busy checking AI output and AI is automating the technical tasks that junior employees once learned from, the path from contributor to leader can quietly narrow. Organisations that invest in leadership development programmes in Chennai can address this directly. Practical sessions on oversight, coaching, and judgement give managers the skills the new role demands, and they protect the time and attention needed to develop the next generation. Skills that are practised together also spread faster than skills each manager works out alone.
Why Structure Matters as Much as Skills
Skills training cannot fix a role that has been designed badly. Deloitte describes a life sciences company that nearly tripled the number of field representatives per manager. Day-to-day operations held up, but coaching and development suffered, and the result was lower engagement and higher turnover. The company eventually found a span of control that protected the coaching role. Flattening without redesign can also create what Deloitte calls shadow leaders and unclear decision rights, because people still need leadership and guidance even when formal titles disappear. Deciding how many people a manager can reasonably support, who holds which decisions, and where AI sits in the workflow is a strategic question. A strategy consultant in Chennai can help leadership test those design choices before they harden into problems.
A Practical Starting Point for Chennai Employers
Begin by finding out what managers actually spend their week doing. Many leaders are surprised by how much time goes on administration and rework. That picture shows where AI could help and where support is most needed. Next, train for oversight and judgement, not just tool features. A session on how to use a particular app can quickly become outdated. A session on how to test AI output, challenge a plausible but weak analysis, and decide when a person should step in stays useful as tools change. Then protect time to learn. The Harvard Business Review researchers found that teams made progress when leaders lowered utilisation targets during the transition and made it easy to share what worked. Finally, reward the behaviours that make adoption stick, such as coaching, documenting good practice, and helping colleagues, rather than measuring delivery alone.
Key Takeaway
AI may change how many managers an organisation needs, but it is also raising the expectations placed on those who remain. The managers who thrive will check work with a sharp eye, make good calls without perfect information, coach people through change, and shape how AI fits into the team's work. These skills can be developed more effectively when organisations treat them as a priority rather than an afterthought. If your organisation wants managers who are ready for the way work is changing, contact us to discuss a practical development plan built around your teams and your goals.
Are your managers prepared to lead teams through AI-driven workplace changes? Contact Exxelo to discuss a practical leadership development programme focused on managerial judgement, coaching and effective team leadership.
Talk to Exxelo


