The democratisation of expertise: how AI is flattening organisational hierarchies for the better

The air in the strategy room carries that familiar mix of tension and curiosity. Around the table, executives trade forecasts and debate market shifts. Then the youngest person in the room, a 24-year-old analyst, leans forward. She’s been tracking live data throughout the discussion. With a few keystrokes, she runs a fresh scenario that challenges the group’s long-held assumptions. Charts shift, the room quiets, and for a moment, hierarchy fades. What matters is the clarity of her insight.

Moments like this are becoming increasingly possible. AI now gives almost anyone the ability to analyse, test, and validate ideas with a depth once reserved for the select few. Tools that used to be locked behind budgets or technical barriers sit on every laptop, narrowing the distance between junior and senior, specialist and generalist. Ideas rise on merit, not title. Leadership, in turn, is evolving: the most effective executives aren’t necessarily the ones who guard decisions, but those who invite them. Influence comes from discernment, the ability to recognise good thinking wherever it appears. Expertise is no longer static or owned; it’s shared and continuously renewed, a living dialogue between people and their tools. And within that exchange, the future of leadership is quietly taking shape.

The old gatekeeping: how hierarchy worked

Organisational structures once defined not just how companies operated, but who had the right to know, decide, and lead. Those foundations are starting to shift.

Organisational structure has always been more than a chart on a wall. It’s the framework that determines how departments interact, how information flows, and how decisions are made. A thoughtful structure brings clarity: people know where they fit, who they collaborate with, and how their work contributes to the organisation’s larger goals. Whether a company has fifty employees or fifty thousand, structure keeps ambition coordinated. For most of corporate history, organisational hierarchy correlated directly with information access. A person’s position in the hierarchy determined what they could access, and access determined what they could influence. The pyramid reinforced itself naturally; junior staff worked with limited data, middle managers aggregated and interpreted, and senior leaders made decisions based on comprehensive intelligence that flowed upward. Each tier justified its existence partly through privileged access to information that lower tiers couldn’t obtain.

At the top, leaders didn’t just have more experience; they also had specialised expertise at their disposal. Customer insights came through commissioned research. Market analysis arrived via consultants. Forecasts were delivered by in-house specialists. The process made sense, given the constraints. Information was expensive to gather and required significant resources. Analysis demanded specialised skills and proprietary tools. The ability to spot patterns only came through years of accumulated experience. And speed rarely determined outcomes when your competitors operated under identical constraints. But that’s no longer the case: data is now abundant and cheap to obtain, and analytical tools are accessible to anyone curious enough to use them. As the pace of business accelerates, the premium has moved from gathering information to making sense of it quickly. The structure remains, but its centre of gravity is moving.

The new reality: what AI changed

AI has made analytical power universal. The difference now lies less in who has access to information and more in who can use it well.

So, how exactly has AI changed things? For starters, even a mere junior marketer can now analyse customer sentiment across millions of reviews; predict campaign performance before a single dollar is spent; segment audiences with precision that would make traditional focus groups look crude, and monitor the moves of competitors the moment they happen. The tools they use are the same ones used by the company’s Chief Marketing Officer – indeed, even the dashboard is identical. In other words, the monopoly on expertise has fractured. AI hasn’t replaced experts so much as made their capabilities available to anyone willing to learn the interface. A curious associate can now do in an afternoon what used to require a team of analysts and a whole month’s work. The gap between what junior staff can accomplish and what senior leaders can accomplish has narrowed dramatically.

What’s interesting is that junior employees often adapt to these tools faster than their managers. They carry fewer assumptions about how analysis should work, which makes them more open to what the technology actually offers. They’re comfortable iterating rapidly because they haven’t spent years perfecting a particular methodology. And they can experiment more freely, since a failed test won’t damage a reputation they’re still building. AI’s growing role inside organisations is also reshaping their structure. As systems automate administrative tasks, individual contributors can operate with greater autonomy. Many of the coordination duties once handled by middle managers, such as allocating work, monitoring performance, or supporting decisions, are now assisted or managed directly by AI. With less need for oversight, organisations can flatten their hierarchies, speeding up information flow and response time.

Flattening the hierarchy

Research by Forrester shows that companies embracing this model have seen a 23% rise in decision-making efficiency and a 37% improvement in conflict resolution, driven by faster access to data and clearer accountability. AI-powered collaboration and knowledge tools amplify this effect, making information visible to literally everyone who needs it. When every employee can consult the same dashboards and models as their managers, the information asymmetry that once justified multiple layers of hierarchy begins to fade. Forward-thinking organisations are already building around this logic. They’re shifting from function-based departments to value-creating teams, each empowered by shared intelligence and supported by adaptive systems, rethinking how work is coordinated and where decisions are made.

AI reduces friction in decision-making along multiple axes. For example, people can simply get the information they need faster. Analysis that would typically take days happens in a matter of minutes, and real-time insights become available across the organisation rather than filtering slowly upward through reporting chains. Whether a company chooses to flatten its hierarchy is still a choice (there’s no universal template), but even modest adoption tends to reduce the need for oversight and speed up operations. When companies combine AI capabilities with looser hierarchical controls, decision quality often improves, particularly in ambiguous situations where multiple paths forward seem reasonable. AI tools can model different scenarios and project likely outcomes, letting decision-makers see further ahead than intuition alone would allow. Instead of committing to a strategy based on incomplete data and executive instinct, teams can make informed calls with a clearer view of probable trajectories.

A new kind of leader

When everyone has access to the same data, leadership stops being about knowing more and starts being about knowing what matters.

The leaders thriving in this new environment aren’t necessarily the ones with the deepest technical expertise or the longest list of facts. Rather, they’re the ones who can tell which insights actually matter. They understand that AI can surface patterns faster than any analyst, but it can’t read the room. It can’t sense timing, nuance, or culture. Leadership now depends on recognising what the model can’t capture, synthesising multiple AI outputs into a coherent direction, and knowing when to trust the analysis and when to question it. Of course, it’s also about knowing which questions to ask.

Under the old hierarchy, curiosity stopped at information. Where traditional hierarchies once revolved around extracting facts, asking what the data revealed, the new reality demands interpretation: how those findings connect to the organisation’s specific goals and circumstances. Where the focus used to be on outlining every possible option, the challenge now lies in determining which path actually reflects the company’s character and long-term vision. And whereas the conversation once centred on predicting competitors’ moves, it now turns to defining what actions align with the organisation’s purpose, regardless of what others might do.

AI can handle the first type of question reasonably well. It can tell you what the data shows, generate a list of viable options, and forecast likely competitor behaviour based on historical patterns. The second question in each pair is where AI starts to falter. Knowing what data means for your specific organisation requires understanding dynamics that tend not to show up in datasets, such as team capabilities, cultural readiness, stakeholder relationships, and the unwritten rules that govern how work actually gets done. Choosing which option aligns with your identity means having clarity about values and purpose that goes beyond optimisation metrics. Deciding what to do independently of competitor actions requires conviction about your own direction. That gap between what AI can answer and what still requires human judgment is where wisdom operates. Leaders who understand that gap, who know which questions to ask AI and which questions to wrestle with themselves, are building the kinds of organisations that can actually use these tools well.

Power dynamics in the AI era

AI is changing what happens inside the room where decisions are made and who gets to shape them.

Strategy sessions used to follow a predictable pattern: senior executives would gather to review reports prepared by analysts, debate the findings, and settle on a direction. Discussion flowed upward; decisions flowed down. The hierarchy defined not only who spoke, but whose perspective carried weight. If you were early in your career, you took notes and only spoke up if someone asked for your opinion. But the modern version of that meeting looks very different. A strategy session begins with a shared definition of the challenge or opportunity. Then, people from across departments bring AI-generated insights and recommendations into the discussion. The senior leader acts more as a guide than an arbiter, shaping the conversation around interpretation, trade-offs, and the context that data alone can’t capture. The outcome is rarely dictated; it evolves through collective reasoning and informed debate. 

AI has also reshaped the power dynamics inside those rooms. Junior analysts now enter discussions equipped with the same tools as their executives. When a model surfaces a pattern that challenges a long-held assumption, it’s no longer dismissed. The senior leader pauses, asks what the system might be missing, and invites the team to test the finding. The analyst isn’t punished for speaking up; she’s valued for improving the outcome. The hierarchy doesn’t disappear, but its edges soften. Authority becomes porous, anchored in learning rather than control.

Blurring the lines between associates and middle managers

Edwige Sacco, who heads Workforce Innovation at KPMG, sees this trend taking hold across industries. “Society is shifting to a skills-based workforce. We are beginning to define someone’s contribution by their skill set, by the value they create, and less so by their role,” she explains. “Because AI can serve as a helper, a mentor, a creative partner, or an information source, anyone can contribute at a higher level, which is blurring the lines between middle manager and associate tasks.”

Middle managers benefit as much as anyone from this shift. “Each problem has a different methodology for fixing it. The expectation used to be that a middle manager was an expert on the methodology. But now, when you bring gen AI into the mix, you can train it to master and explain the methodology to anyone,” adds Sacco. “People can read about the methodology, ask gen AI about it, and come to a strategy meeting already knowledgeable. This frees middle managers to execute on the methodology by speaking directly to clients. They’re spending less time learning and teaching the methodology, and more time linking the methodology to the client’s problem.”

As this shift deepens, promotion criteria are changing too. Progression is no longer awarded to those who have simply accumulated experience or relationships, but to those who consistently exercise sound judgment. A junior employee who synthesises complex inputs and makes good calls, even when they challenge AI’s findings, advances faster than a senior one who simply echoes machine outputs. In this process, expertise still matters deeply. But authority now flows from how well people use it, through discernment, collaboration, and the courage to question both the data and themselves.

The growing pains of a flatter world

Flattening hierarchies sound empowering in theory, but for both leaders and juniors, the adjustment has been anything but simple.

The transition hasn’t been easy for anyone. Indeed, for senior leaders, it often feels personal. Many have spent decades refining their expertise and mastering the art of being the most informed person in the room. Now, even junior colleagues can simply pull up a dashboard and challenge their assumptions, with data to prove it. So, when the advantage they used to have disappears, many leaders find themselves questioning what value they actually provide. If anyone can generate insights, what exactly justifies the corner office? This kind of epiphany can take a serious toll on a person’s ego. KPMG’s Edwige Sacco has experienced this tension up close. “We’ve made great strides, but we still have a percentage of people who are resistant,” she says. “I tell them that AI can’t do one of the most critical aspects of their job. They know how to speak to our clients. They can read between the lines, help others articulate a problem, and explain what they need to solve it. These are human-centric conversations AI can’t have today.”

Junior employees face their own set of challenges. Access to sophisticated tools can breed overconfidence; being able to run an analysis doesn’t automatically mean knowing how to interpret it wisely. AI may be able to surface trends, but it can’t explain the organisational history behind a decision or the unspoken politics that shape its outcome. And because access creates expectation, younger professionals are now under pressure to perform with a sophistication that is well beyond their tenure. While they may have the information, they can’t yet spot patterns in it, sometimes that comes from years of trial and error. They lack the intuition built from past cycles, the timing that separates a good idea from a well-timed one, and the resilience that comes from seeing what failure teaches. AI can’t accelerate that kind of learning. Experience still matters, of course, but it’s just no longer the whole story.

Another challenge is that some hierarchies resist flattening more than others. Regulated industries move more slowly because specialised compliance knowledge still carries real weight. In pharmaceuticals or financial services, for instance, getting decisions wrong can trigger consequences serious enough that organisations hesitate to democratise decision-making too quickly. Risk-averse cultures face similar constraints. When mistakes lead to significant damage, including legal liability, safety failures, and irreversible reputational harm, companies tend to keep critical decisions concentrated among people with demonstrated judgment. Flattening of the hierarchy, then, isn’t a uniform outcome. It’s a spectrum, one shaped by context, risk tolerance, and the willingness of leaders to let go of control. While the tools make it possible, it’s still the people who decide how far to go.

What great leaders are doing

As hierarchies flatten, great leaders are redefining what it means to lead by teaching judgment, offering learning opportunities, and building teams around intelligence, not titles.

So, what exactly separates a great leader from others in this flat new world? Many are pulling back the curtain on how they actually make decisions. For years, much of executive judgment happened behind closed doors or got filtered through management layers before anyone else saw it. When AI gives junior employees access to sophisticated analysis and recommendations, the value leaders provide shifts toward teaching judgment itself, and you teach judgment by making your thinking visible. That means articulating why you’re leaning one direction over another, weighing trade-offs out loud during meetings, and explaining what factors tipped a difficult decision. Transparency here does more than build trust: it accelerates how quickly less experienced employees develop sound judgment by letting them watch the process unfold in real time.

Traditionally, that kind of learning unfolded slowly. Wisdom was absorbed through years of entry-level work and apprenticeship under experienced mentors. But with AI now handling many of the tasks that once served as training grounds, such as data gathering, analysis, and drafting reports, younger employees risk missing the gradual, experience-based growth that teaches judgment. The best leaders recognise this and build new ways to replace it and create deliberate practice for wisdom. Some design “stretch” assignments that expose employees to ambiguity in controlled settings. Others rotate team members across functions so they can see how decisions connect across the organisation. Increasingly, companies are turning to AI-powered simulations to replicate complex scenarios, allowing emerging leaders to test options, make mistakes, and see the consequences unfold. This kind of risk-free practice helps people quickly build the intuition that once took years to develop.

The best leaders are also rethinking how teams are built. Instead of stacking roles by seniority, they’re assembling what you might call ‘intelligence teams’ – that is, groups organised around complementary capabilities rather than hierarchy. A typical team might combine a junior employee fluent in AI tools, a mid-level manager who understands operations, and a senior leader who brings strategic judgment. Together, they form a single decision-making unit, each providing something the others can’t. Some organisations have even gone further, embedding AI directly as a team participant. They assign someone the role of AI orchestrator, a person responsible for managing the system’s inputs and outputs as carefully as they would mentor a talented colleague. In these environments, collaboration includes both humans and machines, with leadership defined less by rank and more by the ability to guide collective intelligence toward clarity.