Building a Platform: Sometimes, You Have to Start Over
Lexandria’s Chief Operating Officer, Josh Sanders, explains how the platform came to be.
Most of us have inherited something that almost works.
A shared drive. A curriculum binder. A platform that once solved a real problem and now mostly creates new ones. You can keep polishing it out of loyalty to the past, or you can admit that the people you serve need something the old thing can no longer become.
That was the choice Lexandria faced in our first months.
When we launched as an organization on July 1, 2025, we inherited something real: tried and true social studies content and a clear picture of what teachers were up against. We also inherited a platform that had done important work in its time but could not evolve into what classrooms needed next. The reading experience felt more like slides than books.
The deeper problem was under the hood. Behind the scenes, there was no real structure: every heading, image, and book chapter lived in one flat, undifferentiated pile, the digital equivalent of a library where every page of every book is filed loose in a single drawer. Updating even small things meant untangling work that had accumulated over years, under many hands, each solving the problem in front of them. No one had done anything wrong. The system had simply outlived its shape. Even the development agency that knew the platform best agreed that rebuilding might serve us better than remodeling.
So the question was never really whether to rebuild. It was how we could afford to. Done the conventional way, it would have taken years and a budget a young nonprofit did not have.
Lexandria exists to reignite the American Spirit in the hearts of young people: to put excellent social studies content in front of teachers in a form that is beautiful, navigable, and usable on a Tuesday afternoon. If the product could not do that well, clinging to the old path was not caution. It was delay.
Three Days
I am not a software developer by training. I came from operations work, the kind of role where you learn how systems fit together, not how to ship a product from a blank page.
In those early weeks, I began experimenting with Cursor, an AI-assisted coding tool that was still relatively new. What it made possible stunned me. Not because a tool can replace judgment (it can't), but because it opened a door that used to stay locked unless you had a large engineering budget and a long timeline.
In mid-August, I went to my co-founders, Richard Lorenc and Sean Malone, with an honest proposal: an experiment, timeboxed to three days, to see how far I could get in building a brand new platform. At the end of that window, we would look at whatever existed and decide whether to keep going.
Those first days were three kinds of work. Some of it was brick by brick: connect the content system, set up the database, make the two talk to each other. Some of it was slow but genuinely fun, like deciding exactly how our Netflix-inspired ebook cards should be sized and where a title should sit on a cover. And some of it had me banging my head against the wall over things that seemed like they should be simple, like getting the library on the screen to match the design we had mocked up. Anyone who has built anything, from a lesson plan to a bookshelf, knows all three kinds.
By the end of the three days, we had a rough prototype. The most important part was a library page.
Sean had already moved our content into a modern system, so real covers and real titles were waiting. When that first library loaded (sparse, unfinished, missing the filters and polish that would come later), you could still see it: a place where great books looked like books again. Not slides. Not a pile of files. A library.
That glimpse was enough. We went for it.
The Tool and the Driver
From that Monday until our October 27 launch, the work was relentless. The tools got better month by month. So did the product. AI assistance accelerated the construction; it did not remove the need for a human driver. Someone still had to decide what mattered for a classroom, what felt worthy of the texts we publish, and what was just noise.
Sometimes literally. When I built the reader's table of contents and the ebook landing pages, the AI kept decorating them: icons, badges, indicators, little bits of visual noise on every line. Helpful in theory, wrong in practice. I removed almost all of it. Nobody teaches a machine that restraint is a feature. Somebody has to be there to say no.
That distinction matters beyond our office. "Built with AI" can sound like a magic trick. It wasn't. It was hours and judgment in service of a simple spine:
Library → ebook landing page → reader.
Find something worth teaching → Open it → Read it in a way that respects the text.
At Lexandria, AI is an assistant, never an author. It helps us build the platform, map content to standards at scale, and sometimes rough out an outline or an early draft. But nothing reaches a student until people who know the subject have shaped it, checked it, and decided it belongs in a classroom.
On October 27, 2025, the new platform launched. In the first week, more than forty teachers registered. Today, nearly two thousand have.
What Teachers Actually Got
Because we rebuilt the foundation, the library could feel like a library. The reading experience could be something students would stay with. Standards alignment and filters could help a teacher find what fits their classroom, not just what happens to surface in a search box.
Here is what that looks like in practice. A high school government teacher has a C3 Framework standard to address this week: explain how the U.S. Constitution establishes a system of government with powers, responsibilities, and limits that have changed over time and remain contested. She searches for that standard on Lexandria. She doesn't get a wall of links to triage. She gets the resources in our library that actually align with it, from a chapter in our government course to a book written for America's 250th on the principles behind the Constitution, each one readable, beautiful, and a click away from her students. Ten minutes of planning instead of an evening of open tabs.
We are still building better ways for teachers to put readings in students' hands quickly and to include more kinds of media, including video, within the same coherent library experience. The details will keep changing. The aim, however, will not — one place where excellent content is easy to find and share, and ready for the classroom.
Teachers already know the cost of tools that almost work. Lexandria's product exists because we decided that "almost" was not good enough for the ideas we ask students to meet.
But there is a deeper reason this story belongs on this blog.
The subject we teach is, at its heart, a story about inheritance. Every generation of Americans receives things it did not build: institutions, ideals, a Constitution whose powers and limits are still being argued two and a half centuries on. And every generation has to decide what to do with what it received. What to preserve. What to repair. And what to have the courage to rebuild so that the essential thing survives.