In 1998, an Oklahoma businessman named Don Greiner made a decision that has been quietly reshaping the conversation about human wellbeing ever since.
Greiner was the president of Grace Living Centers, a chain of nursing homes in Oklahoma. He believed that the residents in his facilities were structurally isolated in ways that were degrading their health and their sense of purpose. The residents ate together. They watched television together. They were cared for competently. And they were, in the specific human sense, alone.
Greiner invested approximately $200,000 of his own money to install two kindergarten classrooms directly inside one of his facilities in Jenks, Oklahoma, in partnership with the Jenks Public Schools West Elementary. The school pays $1 per year in rent.
The design was deliberate. The classroom doors had a six-inch gap at the top so residents in wheelchairs passing by could hear the children. The beauty salon was designed with large windows opening onto the classroom, so residents getting their hair done watched children learning to read. The children ate lunch in the hallway, sitting on the floor, so residents could see them and, over time, know them by name.
The program that emerged was called Reading Buddies — or, in the residents’ own preferred name, Grandfriends. Approximately 40-60 children in kindergarten. Approximately 110 residents with an average age of 81. The principal Suzanne Lair supervised the classroom side. The nursing home director supervised the resident side.
What Sir Ken Robinson claimed happened next
Sir Ken Robinson referenced the Grace Living Center in multiple talks, arguing that the outcomes were dramatic. Children reading above grade level. Residents living longer. Residents reducing their medications. Both populations reporting substantially higher wellbeing.
The educational outcomes for the children are the most robustly documented. Reading Buddies participants, across multiple years, showed measurable literacy and social development advantages over peers in the standard elementary school setting.
The health outcomes Robinson attributed to the residents are less rigorously documented. Advocacy framing has sometimes exceeded what the underlying peer-reviewed literature supports. I flag this because I want the argument in this piece to rest on the reliably documented parts, not on the aspirational parts.
What is well-documented: the children’s literacy and social outcomes improved. The residents’ subjective wellbeing improved. The staff reported the environment was transformed. The program has been replicated across multiple sites. The Greiner investment produced durable value for both populations.
The reframe
The reason this story matters in 2026 has nothing to do with nursing homes or kindergarten classrooms.
The reason it matters is that it names, precisely, what is missing from almost every AI intervention currently being sold as a solution to human loneliness, human learning, and human wellbeing.
A specific example. The current AI companion app market — apps designed to provide human-like conversation to lonely users, marketed to elderly populations, adolescents, and isolated adults — is projected to become a multibillion-dollar category over the next five years.
Every one of these apps is a symptom patch. The user is lonely. The app provides a simulation of companionship. The loneliness is not addressed. It is medicated.
Greiner’s intervention was different in kind. He did not build a companionship simulator. He redesigned the physical infrastructure so that the isolated population was structurally embedded with a population that would provide, as a natural byproduct of the redesigned environment, actual human connection.
The design was expensive upfront. Two hundred thousand dollars. It required cross-institutional coordination between a nursing home and a public school. It required physical construction. It required years of ongoing operational commitment.
What it produced was durable structural change. What the app produces is a transient behavioral simulation.
The general principle
The principle Greiner’s intervention illustrates has broad application. For any human problem that involves structural conditions — isolation, disengagement, underperformance, disconnection — there are two possible response categories.
Symptom patching. Deploy a tool that provides a transient simulation of the missing condition. Cheap upfront. Requires no institutional change. Produces measurable but shallow engagement. Does not address the underlying structural problem.
Infrastructure redesign. Modify the underlying structural conditions so the problem stops being produced. Expensive upfront. Requires cross-institutional coordination. Produces durable change. Addresses the underlying structural problem.
The current AI moment has produced an enormous supply of symptom-patching tools. Some are genuinely useful. Almost all of them are being deployed in situations that would benefit dramatically more from infrastructure redesign.
Three practical questions
One: for each AI intervention your organization is deploying, is it symptom patching or infrastructure redesign? The distinction is not always obvious. But for every intervention, one of the two describes what you are actually doing. Naming it correctly changes the evaluation.
Two: which of your organizational problems, if honestly diagnosed, would be better addressed by infrastructure redesign than by AI symptom patching? Isolated employees. Disengaged customers. Underperforming teams. Almost every one of these has a Greiner-style solution that would produce durable change — and an AI-tool version that would produce a transient simulation.
Three: what is the equivalent of the six-inch gap over the classroom door in your organization? Greiner’s design was full of small, deliberate architectural choices that produced connection as a natural byproduct. What are the equivalent small architectural choices in your workplace that would produce connection, engagement, or learning as a natural byproduct of the redesigned environment?
The closing thought
Don Greiner is not a household name. His $200,000 investment in 1998 did not produce any of the returns venture capital would consider significant. He was not celebrated in the business press. He was not the subject of a TED talk.
He did produce durable structural change in the lives of hundreds of children and residents across multiple decades of operation. He did produce a demonstrated model that has been replicated at other sites. He did produce the story that Sir Ken Robinson used, in his own imperfect way, to point at what the technology-solutionism paradigm was missing.
In 2026, the AI companion app companies will produce enormous financial returns and enormous marketing budgets and enormous celebrity endorsements. They will not solve loneliness. They will medicate it at scale.
The Greiners of 2026 — the executives who invest in cross-institutional infrastructure redesign rather than in symptom-patching apps — will not be celebrated in the same way. They will produce dramatically better outcomes.
The choice, for every senior leader reading this, is which of the two categories your AI portfolio is currently in.
The world has changed. The leaders who notice will be the ones the next decade is built around.
