When Institutions Fail at Service, They Start Monetizing the Breakdown

Jun 16, 2026

10 min read

4 views

84%

The strange new business model of modern institutions

What happens when an institution stops doing the thing it was built to do, but keeps getting paid anyway?

That is the unsettling pattern hiding beneath both academic publishing and FAFSA processing. In one case, publishers turn publicly funded research into a private asset, then sell it back to the universities that paid for it in the first place. In the other, the federal aid system struggles to process student applications, so the solution becomes more layers of support, more consultants, more concierge services, more temporary fixes. In both cases, the institution does not disappear. It mutates. It becomes a tollbooth around a public function.

That is the deeper connection: when systems become hard to use, opaque, or broken, they do not necessarily get repaired. They often get wrapped in a secondary market of intermediaries, licenses, workarounds, and support services. The failure itself becomes productive for someone.

This is not just an administrative annoyance. It is a moral and economic shift. A broken system creates demand for translators, technicians, and gatekeepers. Those who can afford help move through. Those who cannot wait in line, lose time, or drop out entirely. And the institution, instead of simplifying the journey, often learns how to monetize the confusion.

How a public function becomes a private toll road

The old story of institutional efficiency is simple: build the system, serve the users, fix the problems. The new story is stranger. A system can be simultaneously essential and unusable, and that unusability can become part of its business model.

Academic publishing is the cleanest example. Research is funded by universities and foundations. Academics do the labor. Publishers acquire the rights. Then universities pay again to access the output. Now a new layer appears: publishers license that same corpus to AI firms, who use it to train models that may eventually produce new writing at scale. The same public knowledge is repeatedly enclosed, repackaged, and rented out.

The FAFSA rollout reveals a similar structure, though through bureaucracy instead of copyright. Students and families need aid, but technical bugs prevent many from submitting forms at all. Colleges cannot fully process aid without data. So the response is not simply to repair the system at the point of failure. Instead, the department adds experts, concierge services, nonprofit staffing, and test data. Necessary? Yes. Adequate? Not really. The system begins to rely on a support layer to compensate for its own breakdown.

This is the crucial pattern: the institution becomes less a provider of a direct service and more a manager of friction.

Think of it like a highway that develops potholes, missing signs, and jammed toll booths. Instead of repaving the road, the operator hires more tow trucks, traffic officers, and roadside assistants. Travel still happens, but only because a second industry springs up to absorb the chaos. The road survives by outsourcing its own failure.

A system that cannot function cleanly often does not collapse. It survives by shifting the burden of competence onto everyone else.

The hidden cost of “support” as a substitute for design

There is a seductive logic to support services. They feel humane, responsive, even generous. If people are struggling, why not deploy experts? Why not create a concierge model? Why not train more staff and patch the holes?

Because support can become a moral alibi for not fixing the structure.

That is the danger in both stories. In academic publishing, one can point to AI licensing and say that the ecosystem is “creating value” from scholarly content. But if researchers are not paid, if universities pay twice, and if access remains gated, then value creation is just a euphemism for extraction. The support narrative hides the enclosure narrative.

In FAFSA, a concierge service sounds like customer care. But if applicants still cannot submit because contributors without a Social Security number cannot create FSA IDs, then a concierge is not a solution. It is a patch for a machine that still does not work. Families do not need a nicer help desk as much as they need the form to function.

This matters because support layers have a psychological effect: they make dysfunction feel managed. A call center, a liaison, a consultant, a temporary fund, a concierge team, these can all create the appearance of responsiveness. But the deeper question is whether the core process has become unintelligible to the people who need it most.

A useful test is this: does the support layer reduce complexity, or does it merely help users endure it?

If it only helps them endure it, the institution has not solved the problem. It has normalized it.

The premium on human labor in a world that wants to erase it

There is another striking overlap between these cases: both reveal how institutions treat human labor as both indispensable and disposable.

Academic publishing depends on scholars writing, reviewing, editing, and interpreting. Yet the economic system rewards publishers, not the researchers whose labor creates the product. Then AI arrives, promising to automate writing itself. The final insult is not only that scholars are unpaid. It is that their style, their accumulated labor, their discipline, may be reduced to output for machines.

The FAFSA crisis has a different labor problem but the same structure. Aid offices, under-resourced campuses, nonprofit partners, and federal staff must absorb the complexity that the system itself produced. Their labor becomes the buffer between policy and reality. Meanwhile, the students and parents at the end of the line are expected to remain patient while the adults in the room repair what should never have been so hard in the first place.

This is what institutional breakdown does: it turns human beings into compensators. The more broken the system, the more it depends on people making judgment calls, explaining exceptions, and improvising around bad design.

The most expensive part of a broken system is not the broken software, the bad workflow, or the opaque policy. It is the invisible labor needed to make all of that look functional.

That labor is usually undervalued because it is difficult to measure. No one sees the hours spent explaining a confusing form to a panicked parent, or the editorial labor behind a scholarly article that will later train a model, or the extra time an aid administrator spends untangling corrupted records. But without that labor, the institution stops moving.

This gives us a broader framework: every complex institution has a choice between investing in clarity or taxing human patience. When it chooses the second path, it can survive for a long time, but only by burning goodwill.

The real threat is not automation, it is normalization of brokenness

It would be easy to read these examples as complaints about AI, bureaucracy, or corporate greed. But the deeper problem is more general. The threat is not that machines or intermediaries will replace humans entirely. The threat is that institutions will learn to function well enough while staying structurally bad.

That is why AI and FAFSA belong in the same sentence. AI licensing in academia and concierge services in financial aid are both forms of institutional adaptation. But adaptation is not the same as improvement. Sometimes it is just a way to preserve revenue and authority while postponing repair.

In the publishing world, AI firms want massive corpora, publishers want licensing revenue, universities want access, researchers want recognition, and readers want knowledge. But the arrangement can become parasitic if the creators of the knowledge are never compensated and the public never gets broader access. The system does not disappear. It evolves into a more complex extraction machine.

In aid administration, the department wants institutions to keep moving, colleges want funding to arrive on time, and students want aid offers before enrollment decisions lock in. So expert deployment and test data become the visible response. But if the underlying technical failures remain, the system teaches everyone to adapt to delay. Soon delay feels normal.

That is the most dangerous outcome: when people begin treating dysfunction as a permanent operating condition.

Once that happens, institutions stop asking, “How do we fix the process?” and start asking, “How do we keep people from noticing the process is broken?”

This is why the role of temporary support should be examined carefully. A patch is legitimate if it buys time for a real fix. It is corrosive if it becomes the fix. The line between the two is often crossed quietly.

A better mental model: institutions as promises, not machines

We usually think of institutions as systems. That is useful, but incomplete. A better model is to think of them as promises.

A university promises that scholarship will be created, preserved, and shared in a way that advances knowledge. An education department promises that students can access aid without being defeated by the application itself. A publisher promises to disseminate research. An aid office promises to translate policy into support. If the machinery is elaborate but the promise fails, the institution has lost legitimacy, even if it is still operational.

This framing changes the question we ask. Instead of asking only whether the process is efficient, we ask whether the promise is still being honored. That means asking who bears the burden when the promise breaks.

If academic publishing now generates revenue from AI training, then the relevant question is not only whether licensing is legal. It is whether the promise to support scholarship has been replaced by a promise to monetize scholarship.

If FAFSA needs a concierge service to function, then the relevant question is not only whether the rollout is being managed. It is whether the promise of accessible aid has been replaced by a promise of managed frustration.

Promises are useful because they expose the moral center of institutions. Machines can be optimized around throughput. Promises must be evaluated by trust.

And trust has a simple rule: once people believe the system is designed around their inconvenience, they stop seeing support as care and start seeing it as containment.

Key Takeaways

Watch for support layers that grow faster than fixes. If concierge services, consultants, or help desks expand while the core problem remains, the institution may be managing breakdown instead of solving it.
Ask who profits from friction. If complexity creates new revenue streams, licensing deals, or staffing contracts, the system may have incentives to preserve, not eliminate, the pain points.
Separate patching from repairing. A patch buys time. A repair reduces future dependence on patching. Use that distinction to judge whether a solution is real.
Measure invisible labor. The people keeping broken systems afloat are often the least rewarded. Identify where human effort is compensating for bad design, and treat that as a structural cost.
Reframe institutions as promises. Efficiency matters, but trust matters more. Ask whether the institution still delivers on the reason it exists.

The final test: does the system need you to forgive it, or to use it?

The deepest connection between these stories is not AI or financial aid. It is a question about modern institutions under stress: when they fail, do they repair themselves, or do they ask everyone else to become more adaptable?

That is the difference between a service and a toll road, between a public good and a managed inconvenience, between an institution that serves and one that merely survives.

The more a system depends on your patience, your expertise, or your willingness to tolerate delay, the more likely it is that the system has shifted its real function. It is no longer just delivering value. It is extracting resilience.

And once you see that, you start noticing it everywhere: in the forms that never quite work, in the articles that will be mined by machines but not fairly rewarded, in the help lines that exist because the product is unusable, in the quiet growth of entire industries built around fixing what should not have broken.

The most important reform is often not a grand redesign. It is this: stop rewarding institutions for making you adapt to their failures. Make them earn the right to be called useful again.