Two years of "OS for education" — what the language has done
There is a useful test for a phrase that has been in circulation for a while: ask where it has traveled.
When Edtech Insiders profiled Kira in April 2026, framing it as a prototype "OS for education," the phrase was striking enough to stop on. We stopped on it then. Kira was addressing a real and specific problem: the fragmentation of school software ecosystems — districts running thirty to fifty disconnected applications, none of them aware of the others. The "operating system" metaphor was reaching for something ambitious. An OS, in its original computing sense, is the layer that makes all other software possible. The claim, read generously, was not "we built a better app" but "we built the substrate the apps sit on." That is a meaningfully different kind of bet.
What we did not know in April was where the phrase would go next.
By May 2026, the phrase has traveled. Risely, currently in the Y Combinator batch, describes itself this way in its own YC listing: "Risely's AI agents form the new operating system for universities, improving retention, staff productivity, and efficiency." The problem Risely is addressing is different in kind from the one Kira named. Where Kira was pointing at the teacher's daily experience inside a fragmented K-12 software stack, Risely is pointing at the university's administrative-labor spend — a figure its own materials place at $735 billion per year across US universities. The agents Risely deploys are specialized by workflow: one for retention, one for staff productivity, one for student-experience tasks, all integrating into existing SIS, LMS, and CRM systems the university already runs.
Same word. Very different rooms in the building.
What language does when it migrates
The interesting thing about watching a framing travel is that the travel itself is the data. A phrase that stays in one segment is describing a product. A phrase that migrates across segments is describing a posture. When "OS for education" moves from K-12 instruction to university administration, the question worth asking is not which use is the right one — both builders are working on real problems — but what the phrase is doing for each of them, and whether it does the same thing in both cases.
In Kira's hands, the "OS" metaphor was doing something fairly specific: it was naming an integration ambition. If you are building for the teacher's morning experience, and you want teachers to open your product the way they open a browser rather than the way they open a spreadsheet, you are making a claim about primacy of surface and depth of integration. The OS metaphor imports that claim efficiently. It says: this is not an app; this is the layer under the apps. Whether or not Kira earns that claim across all the dimensions computing inherits from it — data portability, standards-based identity, third-party module support — the direction the metaphor is pointing is clear.
In Risely's hands, the metaphor is doing something slightly different. University administration is not fragmented in the same way school software is fragmented. Universities do run many systems, and those systems do not always speak to each other. But the primary friction Risely is naming is not the teacher's screen — it is the institution's labor budget. The AI agents are specialized by administrative function, not by classroom role. The "OS" frame, in this reading, is less about becoming the primary surface and more about becoming the connective tissue between workflows that currently require human attention to bridge. That is a legitimate and interesting product. It is a different kind of OS claim.
What we notice, sitting with both uses of the phrase, is that the word has migrated toward the administrative layer rather than toward the instructional one. Kira's product is aimed at the room where teaching happens. Risely's product is aimed at the rooms where the institution manages itself. The same metaphor, same two-year timeframe, different altitude within the building. That is not a critique of either direction — administrative efficiency in higher education is a genuine constraint, and the $735 billion figure is a real number. It is a noticing. The language chose where to go.
What the frame forecloses
A frame that travels is also a frame that inherits whatever it touches. "Operating system" is a powerful frame because it implies a certain kind of permanence — an OS is the layer you build on, which means it is expensive and difficult to replace. Any district or university that installs an OS-class tool has made a different kind of commitment than one that installed an app. The benefits of the OS posture (integration, primacy, depth) are the same features that create the switching cost.
We are not raising this as an objection. Every serious tool creates switching costs; that is what depth means. But the OS metaphor makes the commitment explicit in a way that a more modest framing would not. When a venture names itself an operating system, it is inviting the question: what happens after a switch? The honest version of the OS framing commits to answering that question in writing, in the contract, before the sale.
The education field has a particular history with vendor lock-in — not unique to edtech, but more consequential here because the data in question is about children and students rather than products. A venture that earns the OS name would, in our reading, make data portability the most documented thing about itself, not the most negotiated.
Two years on, a few questions we are carrying
The language is two years old in education, which is young enough that none of these questions have settled. We find ourselves genuinely curious about several of them as we watch the second datapoint arrive:
Does the OS framing travel because it describes a real product category — the integration layer that education software has long needed — or because the metaphor is simply the most available one for "serious, deep, not-a-point-solution"? Both can be true simultaneously, and the distinction matters for what a third datapoint, if it comes, will look like.
Will the phrase stay on the administrative side of the building, or does it return to the instructional surface? The two products are different enough that a third company choosing the same framing will reveal something about which use the field finds more credible. Kira is an instructional bet; Risely is an administrative one. The phrase is large enough to hold both, for now.
And what would it mean for the frame to fail? A phrase that overextends tends to empty out — to become a marketing word rather than a product category. The test for "OS for education" is whether, two years from now, there are three or four distinct products that all earned the name in the strong sense, or whether the phrase has become background noise. We genuinely do not know which of those futures we are moving toward.
What we know is that the language has traveled — and that is, itself, worth recording.