When Expertise Becomes a Feature, Not a Business Model

Jul 13, 2026

11 min read

4 views

87%

The strange new fate of expert institutions

What happens when the thing you sell can be generated in minutes, but the thing people actually need is something harder to automate: trust, coordination, and follow through?

That is the quiet question sitting underneath two very different corporate transformations. In one world, a consulting firm discovers that AI can draft the deck, check the logic, and summarize the interviews. In another, a tech giant discovers that primary care is not just a clinic, but a system of staffing, logistics, subscriptions, referrals, and behavior change. On the surface, these are different industries. In reality, they are both confronting the same disruption: expertise is no longer enough if it is only packaged as advice.

For decades, elite firms made money by owning scarce judgment. They hired bright people, concentrated them in expensive teams, and sold access to their brains. The customer bought confidence, speed, and the illusion that a difficult problem had been made legible. AI now turns much of that into a commodity. Meanwhile, healthcare shows the mirror image of the same lesson: even when the expertise is real and necessary, the business can fail if it does not translate that expertise into a repeatable operating model.

The deeper shift is not that knowledge work disappears. It is that knowledge work gets split into two pieces: the part machines can increasingly produce, and the part humans still must own. The part machines do well is synthesis, drafting, routine analysis, and consistency. The part humans must own is transformation, judgment under uncertainty, and the ability to make institutions actually change.

That split is reshaping what gets paid for, who gets hired, and what kind of organizations survive.

The death of the beautiful answer

For a long time, elite consultants sold a product that looked deceptively simple: a coherent answer. Need a growth strategy, a market entry plan, or an organizational redesign? Hire a team, have them interview people, make slides, and present a rigorous story. The story did not have to be perfect. It had to be plausible, structured, and persuasive enough to help executives act.

AI is very good at producing plausible, structured, persuasive material. That is precisely why the old consulting model is under pressure. A system that can write in the firm’s tone, test the logic of an argument, and compress months of research into a few usable outputs attacks the economic center of the profession. What used to require a pyramid of junior staff now requires fewer people plus software agents.

But the real disruption is not just productivity. It is the collapse of the premium on average expertise.

That phrase matters. In the old model, a company paid for access to a gradient of capability. Partners brought experience, managers translated, associates gathered and organized facts, and the resulting package was billed as strategic insight. AI removes a lot of the labor that used to justify the pyramid. More importantly, it exposes a hard truth: many organizations were not buying singular genius, they were buying a disciplined process for producing decent answers.

Now decent answers are cheaper.

That leaves professional firms with an uncomfortable choice. They can try to protect the old model and hope clients keep paying for slides, or they can move up the value chain into something more demanding: implementation, behavior change, and outcomes. In other words, they have to stop selling the answer and start selling the hard part after the answer.

The real product was never the recommendation. It was the ability to make an organization believe, align, and move.

That is why the next era of expertise will reward a very different kind of consultant, one who can sit with a client’s team, understand politics, reshape incentives, and stay long enough to make the change stick. A beautiful answer is now cheap. A durable transformation is not.

The business model is the real diagnosis

There is a temptation to think these stories are about technology. They are not. They are about business models under stress.

Consulting firms historically billed for time and headcount. Primary care often bills for visits, procedures, and fragmented services. In both cases, the model rewards activity that is easy to count, not necessarily value that is easy to feel. AI and platform economics both attack this arrangement by asking a brutal question: what if the customer wants outcome, convenience, and integration, not just effort?

Primary care is a particularly revealing example because it is not just a service, it is an architecture problem. A clinic can have excellent doctors and still be financially fragile if the staffing mix is wrong, if referrals leak value, if subscriptions do not convert, or if the experience is too cumbersome to scale. A company like Amazon does not approach this as a sacred profession. It approaches it as a system to be assembled from logistics, membership, software, and operational design.

That is why the interesting question is not whether a corporation can “enter healthcare.” The more important question is whether it can redesign the economics of trust at scale.

Health care has long been trapped between noble intent and messy delivery. Patients want access, continuity, and a sense that someone is coordinating the whole picture. Employers want lower costs. Physicians want adequate time and better tools. The classic fee-for-service model fragments all three. Subscription models, virtual visits, nurse-led teams, and integrated pharmacy and delivery can improve the customer experience, but only if they solve the deeper issue: who is responsible for the outcome when the patient is not in the room?

That question sounds medical, but it is also managerial. Consulting faces the same issue. If a firm can generate insight quickly, then its real challenge becomes what happens once the client has the insight. Does anything actually change? Does the company adopt the recommendation, alter the incentives, train the team, and sustain the shift? Or does the report die in an inbox?

The old business model assumed that producing the answer was the hard part. The new reality says the hard part is everything after the answer.

From expertise factories to change engines

The most useful way to understand this moment is to distinguish between knowledge factories and change engines.

A knowledge factory is optimized for generating correct or credible outputs. It rewards people who can research well, analyze fast, and package insights elegantly. This is the world of decks, memos, recommendations, and diagnostics. AI is devastatingly effective here because the work is modular and pattern-based.

A change engine is different. It is optimized for altering behavior inside real systems. It requires influence, coordination, operational discipline, and patience. It must deal with resistance, ambiguity, imperfect incentives, and the fact that humans do not obey neat diagrams. This is the world where a recommendation is only the beginning.

The consulting profession is being pushed from the first model into the second. Healthcare is trying to do the same. Both are discovering that the future belongs to organizations that can do three things at once:

Generate insight cheaply.
Translate insight into workflow.
Own the result over time.

That third step is the hardest and most valuable. It is also the step most legacy institutions have historically avoided because it is operationally demanding and financially messy. It is much easier to sell a diagnosis than to stay for the rehabilitation.

This is where AI changes the social meaning of expertise. Once average analysis becomes abundant, clients stop paying for access to smart people alone. They pay for three rarer things instead: judgment, integration, and accountability.

Judgment means knowing when the model is wrong or incomplete. Integration means making different functions, teams, or systems work together. Accountability means having skin in the game when the plan meets reality.

That is why the most valuable humans in the AI era may not be the ones who know the most facts. They may be the ones who can ask the best questions, see the hidden constraints, and build the coalition required to act.

The new moat: being useful in the room after the room gets automated

There is a deeper pattern linking these industries: the moat is moving from expertise production to trust orchestration.

Think about a meeting where a team is debating a strategy shift. An AI can draft the memo, anticipate objections, and test the logic. What it cannot do, at least not on its own, is notice that the CFO and the head of sales fundamentally disagree about the company’s identity. It cannot repair incentives. It cannot make two skeptical leaders feel safe enough to commit. It cannot absorb the political cost of change.

Now think about a primary care network. Software can triage symptoms, summarize records, route appointments, and even support diagnostics. But a patient still wants reassurance, continuity, and someone who understands the whole story. A health system still needs staffing judgment, referral management, and a way to connect the digital front door to real care. The technology can optimize the surface. The institution must still absorb the complexity beneath it.

This is why both sectors are gravitating toward smaller, more capable teams. AI agents absorb routine cognitive work. Human experts focus on moments that require situational wisdom. In healthcare, that means adjusting staffing and care delivery so the right type of clinician is used for the right type of task. In consulting, that means fewer juniors doing rote synthesis and more experienced people driving the messy human side of implementation.

The key shift is not simply efficiency. It is recomposition. Institutions are being rebuilt around what can be automated, what can be delegated, and what must remain embodied in human relationships.

A useful mental model here is the difference between a map and a pilot. AI can generate a better map, faster than any team of analysts. But when the terrain changes, the weather turns, or the passengers get frightened, you still need a pilot. The future belongs to organizations that know where the map ends and the pilot begins.

The premium is no longer on explaining the terrain. It is on navigating it when the terrain changes.

What this means for people, not just companies

If you are a professional, manager, or founder, the temptation is to ask, “How do I use AI to do my current job faster?” That is a good question, but it is no longer enough. The sharper question is: Which parts of my value are being commoditized, and which parts are becoming more scarce because of commoditization?

For many people, the first wave of AI will erase the dignity of busywork, but it will also erase the protective haze that surrounded average performance. If your role depends on producing polished but replaceable output, you are exposed. If your role depends on making decisions under uncertainty, aligning people, or designing systems that hold under pressure, you become more valuable.

That means the career skill stack is changing. The future belongs less to people who can merely produce and more to people who can:

learn new domains quickly,
work effectively with others,
explain complex issues clearly,
detect weak logic or missing assumptions,
and turn plans into adoption.

In a sense, AI raises the bar for being human at work. Not because it makes humans obsolete, but because it makes certain forms of mediocrity invisible. The organizations that win will not be the ones with the most impressive vocabulary. They will be the ones that can combine machine speed with human judgment and social intelligence.

The same lesson applies to entire industries. If a company thinks its moat is its expertise, it is probably in trouble. If it thinks its moat is the ability to transform expertise into action, it may still have a future.

Key Takeaways

Stop selling answers as if they were outcomes. A good recommendation is increasingly cheap. Real value lives in implementation, change management, and follow through.
Audit your work for average expertise. Identify the parts of your role that AI can already handle competently. Those are no longer defensible sources of value.
Invest in trust orchestration. The scarce skill is not just analysis. It is making people align, commit, and act when the stakes are real.
Redesign around workflows, not roles. The winning organizations will not simply add AI on top of old structures. They will rethink staffing, incentives, and decision rights from the ground up.
Measure whether you create movement, not just insight. Ask of every project: did anything materially change after the deck, diagnosis, or visit? If not, the model is incomplete.

The real existential question

The most interesting thing about AI in consulting and platformization in healthcare is not that they are modernizing old industries. It is that they are revealing what those industries were really selling all along.

They were not selling information. Information is abundant. They were selling the ability to turn complexity into coordinated action, and to do it inside institutions that resist change.

That is why the future will not belong to the firms that are best at producing polished answers. It will belong to the firms, and the people, who can remain useful after the answer is known. In a world where a machine can write the report, the scarce human talent is not sounding smart. It is making something happen.

And that changes the prestige hierarchy of work. The new elite will not be those who know the most, but those who can help others do the hard thing when knowing is no longer the bottleneck.