Our story
Districts have never been short on information.
They’re short of the thing that turns information into a decision everyone understands the same way. This is why we started, what we think is genuinely hard about it, and where the name came from.
Chapter one
Every answer already exists, but it lives in six different places.
A school district runs a dozen separate computer systems. Attendance in one. Test scores in another. The teachers’ union contract is a PDF someone emails around. Special education records are somewhere else entirely.
A superintendent asks how many teachers will we need to hire next year? Answering it means pulling from five systems and reading a contract. Today that takes a person a week.
Someone asks a plain question in a Tuesday cabinet meeting. Why did sixth-grade math slip? Everyone in the room knows the answer is knowable. Attendance is taken every period. Diagnostics run three times a year. Grades, enrollment, staffing, intervention logs, board policy and an LCAP with metrics you promised the community you would hit. It’s all there, and it has all been there the whole time.
So it becomes a data request. It goes to the two or three people who can actually join those systems, who are the same two or three people running state reporting and the assessment calendar. It comes back three weeks later as a district average.
By then the budget conversation has happened. Somebody has already proposed buying a new curriculum.
Nobody in that story is bad at their job. The analyst isn’t slow, the superintendent isn’t impatient and the vendors aren’t lying. The systems were simply never built to be asked a question together, and the cost of that lands on a sixth grader who needed something different in October.
Quilt should answer it in a minute, and be trustworthy enough that the answer survives a board meeting.
The scarce resource in a school district isn’t data. It’s the ability to act on it while it still matters.
Chapter two
And getting the answer is only half the problem.
The other half is that a district isn’t one organization. It’s a dozen of them holding hands.
Academic services has a strategic plan. Business services has a budget built on different assumptions. Special education runs on a SELPA handbook. Every school has an improvement plan, every department has priorities, the board has goals, and thousands of decisions get made every week by people who each have a real and partial view.
All of them are pointed at the same children. Almost none of them are working from the same page.
That isn’t a failure of care. It’s what happens when the work is spread across more systems than any one person can hold in their head. And it’s why two people can look at the same district and honestly disagree about how it’s doing.
We think this is the part AI is actually good for, and it isn’t the part everyone is building.
Not a smarter search box. A district where the goals, the plans, the policy, the data and the institutional memory are close enough at hand that they can inform the work while the work is happening.
A principal writing a school plan can see how it lines up with what the district already committed to, before the plan is submitted, not after.
A department leader weighing a new program can see what evidence already exists in their own district, including the evidence against it.
A superintendent can see where the work is aligned, where it’s duplicated, and where the district has quietly drifted from what it said it would do.
A teacher can see not just what they’re being asked to do, but why it matters and where it fits.
The best districts we’ve worked with aren’t the most data-driven ones. They’re the coherent ones, where people at every level can say where the district is going, why, and what their part of it is.
Chapter three
Which is how we ended up naming a software company after a blanket.
Every other name in this category is a machine word. Platform, layer, hub, engine, brain. Each one puts the software in the middle of the sentence and the school somewhere off to the side.
A quilt is the opposite kind of object. It’s made by people, out of what they already had, for someone specific. Four things are true of every quilt ever made, and each one turned out to be a decision about how to build this.
- It’s made from what you already own.
- No quilt has ever begun with new cloth. It begins with what’s already in the house. You aren’t short of data, so Quilt doesn’t ask you to replace your student information system, re-enter anything, or buy a new source of truth. It works with the cloth you have.
- The seams show.
- You can always see where one piece ends and the next begins. That isn’t a flaw in the making, it’s the construction. Every number Quilt gives you arrives with the system it came from and the moment it was read. Software that blends everything into one smooth surface is hiding its joins.
- It’s made by many hands.
- Quilts were made around a frame, by people who each worked their own corner and none of whom did the whole thing alone. Your district already works this way. Quilt is built so each person sees only their corner, and the whole thing still holds together.
- It’s made to keep someone warm.
- Nobody quilts in the abstract. A quilt is made for a particular person, and usually that person is a child. That’s the test we apply to our own roadmap. If we can’t draw the line from a feature to a student, the feature is for us and not for them.
So it isn’t a platform, a layer or a brain for your district. It’s stitching. And stitching is never the point of a quilt. It’s just what lets the quilt be one thing instead of a pile of pieces.
Connect the pieces, strengthen the whole.
Chapter four
We’re the people who built the systems this now reads from.
That isn’t a boast. Between us we’ve spent decades shipping the assessment, curriculum and analytics products that districts already use, which means we know exactly how this category gets sold, and exactly how it gets abandoned.
We’ve built dashboards that nobody opened twice. We’ve watched careful work get bought in the spring and quietly stop being used by November, not because it was wrong but because using it required a person who didn’t have the hour. We’ve sat in the meeting where a district was told its data was now unified and could tell, from the room, that nothing had changed.
Districts have earned every bit of their skepticism about education technology, and the useful response isn’t to pretend we’re one of you. It’s to build the part we kept watching everyone skip, the authorization, the connectors, the citations, the refusals. The parts that decide whether a person trusts an answer enough to act on it in front of a board.
We’re early, and we would rather say that than imply otherwise. What we can promise is that the boundaries on this site are the same ones we’ll put in your agreement.
How we build
Six commitments we would rather lose a deal than break.
The boring layer is the product
Anyone can wire up a chat box. We spend our time on authorization, connectors, temporal correctness and evals, the parts nobody demos.
Say what it doesn’t know
A confident wrong answer in a board meeting costs a district more than no answer. Quilt is built to say “insufficient evidence” out loud.
Humans decide about children
Software can surface evidence, show policy and organize a decision. It should never make one about a child’s placement, discipline or services.
Constraints in the contract
A promise on a website can be revised quietly. A term in a data privacy agreement can’t. If it matters, we’ll write it down.
Vendor-neutral by design
No district should be re-procuring because we bet on one AI lab. The abstraction is ours. The models are replaceable.
Don’t overstate
We’re early and we say so. No invented customer counts, no certifications we don’t hold, no capabilities that exist only in a demo.
Bring us the question you gave up on.
Thirty minutes, and bring the one that turned into a data request and never came back. We’ll tell you honestly whether we can help.