THE CONVERSATIONS THAT SCALE AXL: Part 1
ANANTHI BALAKRISHNAN
Ananthi leads program implementation in the AXL team at EkStep Foundation
What Actually Scales
When a state government adopts AXL, what scales afterward is not the software. A tool, once built, doesn't need to be rebuilt for the thousandth school, the way it was for the first. What has to happen multiple times over, without shortcuts, is a conversation — the right one, with the right person, at the right moment. This series is about that cascade: what it looks like, who it involves, and why it is the actual determinant of whether AXL works at scale.
The Cascade,
Stage by Stage
Getting a state to commit takes very few conversations, but they carry enormous weight. Two to three conversations, with the academic and administrative heads of the state education departments, decide whether a state signs on at all. These conversations rest on the trust and understanding that technology can play an instrumental role in improving outcomes. What's being decided is whether a government is willing to commit administratively and politically to a new program. The outcome, if it goes well, is a signed MoU. If it doesn't, nothing downstream ever happens.
Turning that commitment into a program takes a different set of conversations — three to four, over roughly two months, with a different set of people: the state program director, the commissioner, and district/block-level officials and teachers. These conversations aren't about whether to do this; they're about how. Which schools go first. How the program design fits the state's priorities and practices. What the rollout sequence looks like. The outcome is a ready school list and a program design specific enough actually to launch.
3.6 million
student interactions
Then the program goes live — and the number of conversations required jumps by orders of magnitude. Once schools begin using AXL, sustaining that usage requires not two conversations or four, but thousands, continuously, for as long as the program runs. A state running AXL across its schools at an early scale of about 1000 schools, and 20000 children, generates roughly 3.6 million student interactions with the platform in a school year. Behind those 3.6 million student sessions sit more than 370,000 human conversations, making sure they happen and mean something.
A teacher checking a dashboard and adjusting a lesson, roughly 300,000 times a year across a state. A head teacher reviewing a week's data with their staff, roughly 60,000 times a year. A block officer reviewing a fortnight's progress, roughly 6,000 times a year, and district and state officials checking in on an ongoing basis above that.
Why the ratio matters
What should stand out in those numbers is not just their size, but their shape. Two conversations get you a signed MoU. Three or four get you a launch-ready program. But sustaining that program at an early scale takes 370,000 conversations a year — a ratio of roughly 1 : 1,000 : 100,000 across the life of an early-state scale partnership. Almost none of the work of scaling AXL happens at the top of that ratio, where a handful of conversations decide whether a state says yes. Almost all of it happens at the bottom, in a rhythm of daily and weekly conversations between people most policy discussions never mention: teachers, head teachers, and block-level coordinators.
This has a direct implication for what “meaningful usage” actually requires as infrastructure. It is not enough to get a government to say yes, or even to design a good rollout plan. Both of those are real achievements, and both are necessary. But they are the smallest part, by volume, of the work. The overwhelming majority of what determines whether AXL succeeds in a state happens in conversations that are individually unglamorous, repeated hundreds of thousands of times, by people who will never sign an MoU or appear in a program review.
Three kinds of conversation, not one
Not every conversation in this cascade is doing the same job, even within a single stage. Broadly, we've found it useful to think of three distinct purposes running underneath the cascade described above:
Commit — conversations that secure a decision to begin, at the state or district level.
Configure — conversations that turn a commitment into a specific, workable plan.
Sustain — the recurring conversations, daily and weekly, that keep a live program functioning at the level of individual teachers and classrooms — and that continuously correct course as the data reveals a stall, a discrepancy, or a plan that needs rethinking. Correction here isn't a separate event triggered occasionally; it's part of the same continuous rhythm that keeps the cadence meaningful rather than mechanical.
The framework holds together because it maps onto something real, not onto a theory: a handful of high-stakes conversations open a door, a slightly larger set of conversations build the room, and then the vast majority of the effort — hundreds of thousands of daily and weekly exchanges between teachers, head teachers, and block officials — keeps that room in continuous use. None of those human conversations touch a child directly. What they do is quietly sustain the 3.6 million interactions a year between a child and AXL that everything else makes possible at this early stage of scale — a number that will need to grow by orders of magnitude again as more states move through Commit and Configure.
