One Lesson, Three Readiness Levels: The Differentiation Problem, Solved

Ask a teacher what they'd do with more time and, near the top of the list, you'll hear some version of differentiate more. Not because they don't know how — they trained for it — but because doing it properly, for a real class of twenty-five students sitting at three or four different readiness levels, means writing three or four versions of tonight's lesson. That's a task with enormous instructional payoff and an equally enormous time cost, which is exactly why it's the first thing sacrificed when the week runs out of hours.

That collision — high value, high cost — is the specific problem Claude for Teachers is built to attack. The framing behind the whole launch is that the best instructional practices are well established, but a teacher's week doesn't allow them. Differentiation is the cleanest example. When the tiered-lesson task drops from an evening's work to a few minutes, the practice stops being an aspiration and becomes something you can actually do every day. This piece is about why that particular task is the right one to solve, and what "solved" looks like when the core content stays consistent across every tier.

Why differentiation is the task worth solving first

Differentiation sits at an unusual intersection. The research on it is strong — meeting students at their readiness level is one of the practices that reliably helps — and yet it's structurally hard to sustain because the cost scales with the diversity of your room. The more your students spread across levels, the more versions you owe them, and the rooms with the widest spread are often the under-resourced ones where teachers already have the least slack. The practice most likely to help the students who need it most is the one hardest to deliver at scale.

That's what makes it the right first target for a teaching skill. Automating a low-value task saves a little time on something that didn't matter much. Automating differentiation collapses the cost of a practice that matters enormously — and it does so exactly where the strain is heaviest. You're not shaving minutes off busywork; you're making a genuinely better instructional practice feasible for a whole class, every day.

The point isn't that AI writes your lessons. It's that a task with real instructional value — one teachers believe in but can't afford — becomes affordable. That changes what a teacher's week can hold.

What "solved" actually means: same destination, different routes

The failure mode of quick differentiation is dilution. It's easy to generate an "easier" worksheet and a "harder" worksheet and declare victory — but if the three versions quietly teach three different things, you haven't differentiated, you've fragmented your class. By the end of the week, your groups have diverged, and reassembling them for the next unit is harder than if you'd never split them.

Real differentiation holds the destination fixed and varies the route. Every student aims at the same core content and the same standard; what changes is the scaffolding, the pacing, and the level of challenge on the way there. The k12-lesson-differentiation skill built into Claude for Teachers is designed around precisely this constraint — it produces below-, at-, and above-level versions while keeping the core content consistent across tiers, and returns a differentiation plan plus per-proficiency student materials. Here's what that looks like in practice for a single lesson:

• Below proficiency — more scaffolding and smaller steps toward the same concept: a worked example before independent practice, a vocabulary support, a partially completed organizer. The idea is unchanged; the path in is gentler.
• At proficiency — the lesson as designed, on grade level, aimed straight at the standard. This is the anchor the other tiers are calibrated against.
• Above proficiency — genuine extension of the same idea: a harder application, a transfer task, a "now justify why" prompt. Not unrelated enrichment busywork — a deeper cut of the core content.
• Specific student needs — targeted adaptations layered on top, for the particular supports individual students require, without changing what everyone is learning.

Read that list and the discipline is obvious: four different left-hand columns, one shared right-hand column. That's differentiation done the way a teacher would defend it in a planning meeting.

From an evening's work to a few minutes

The reason this used to eat evenings is that each version is real work — you're re-scaffolding, re-writing prompts, re-pacing, and producing the actual student-facing materials for each tier, not just a note-to-self about what you'd change. Do that honestly for four tiers and you've written most of four lessons.

The skill compresses that. You start from a lesson you already have, and it drafts the tiered versions and the accompanying student materials for you to review. Your time shifts from production to judgment: you read three drafts and decide whether the scaffolds are right and the challenge is real, rather than building all three from scratch. That's the shape of the whole Claude for Teachers approach — Claude does the drafting, you keep the professional call — and differentiation is where it pays off most, because the production cost was what made the practice unaffordable in the first place.

It's worth being precise about what does and doesn't change. The skill doesn't decide who belongs in which tier — that's your read of your students, informed by your diagnostics and your knowledge of the room. It doesn't replace your judgment about whether a scaffold is pedagogically sound. What it removes is the mechanical cost of producing every version. The thinking stays yours; the typing doesn't.

What changes when the cost disappears

When differentiation stops being an evening project, its role in the week changes. It becomes something you can do routinely rather than reserve for observation days. You can differentiate a lesson you decided to reteach that morning, because the turnaround is minutes. You can adjust tiers as your read of the class evolves, because re-running the task is cheap. The practice becomes responsive instead of pre-planned — which is closer to how differentiation is supposed to work anyway.

That's the real prize. Not a faster worksheet generator, but a shift in what a teacher can sustain. The best instructional practices don't fail because teachers reject them; they fail because the week can't fund them. Take the funding cost off the most valuable one, hold the core content constant so the class stays coherent, and you've done something more useful than saving time — you've made a better classroom possible on an ordinary Tuesday.

To go a level deeper on the mechanics, read the inside look at the differentiation skill; for the step-by-step version of the task, see the differentiate-for-every-learner workflow. And if you want to try adjacent tooling now, education agent skills and lesson-plan-studio live in the AI tutoring category, alongside more builder insights.

Part of the Claude for Teachers series. Related: Inside k12-lesson-differentiation · Differentiate for Every Learner: A Workflow. Browse AI tutoring skills or more builder insights.