Differentiate for Every Learner in the Room: A Workflow
How the Claude for Teachers differentiation workflow turns one lesson into tiered versions — below, at, and above proficiency — with a differentiation plan and per-level student materials that keep the core content consistent.
Ask any teacher what they'd do with an extra ten hours a week and differentiation comes up fast. The class in front of you is never one class. It's a student two grade levels behind on reading, a student who finished the challenge problem before you stopped talking, and a wide middle that needs the lesson pitched squarely at them. Everyone deserves a version aimed at where they actually are. Almost nobody has time to write three.
That time cost is the whole reason differentiation is the practice teachers most believe in and least manage to do. It's not a knowledge gap — teachers know how to scaffold and how to stretch. It's an hours gap. The Claude for Teachers "differentiate for every learner" workflow is built precisely to close it: hand it a lesson, and it produces a differentiation plan plus personalized student-facing materials for each proficiency level, keeping the core content consistent across all of them. This is how the workflow runs, and how to run it well.
Start with a lesson, not a blank page
Differentiation is an adaptation, not a fresh creation, and the workflow reflects that. You begin with an existing lesson — one you wrote, one from your curriculum, or one you just generated with the lesson-planning workflow. Claude's job is to take that single lesson and produce tiered versions of it, not to invent three unrelated lessons that happen to share a topic.
That distinction is the design's most important move. The open-source k12-lesson-differentiation skill behind the workflow adapts a lesson into versions pitched below, at, and above proficiency level, and for specific student needs — but it holds the core content steady across tiers. Every student is learning the same standard, engaging the same essential idea, headed for the same destination. What changes is the on-ramp: the scaffolding, the reading load, the number of steps made explicit, the amount of productive struggle. That's real differentiation. The alternative — easier kids get a lighter topic, advanced kids get a harder one — is just tracking with extra steps, and it's the thing this workflow is careful not to do.
What actually changes across the tiers
It helps to see what "same content, different access" means concretely. Across the three readiness levels, the target standard is fixed; the support scales.
| Element | Below proficiency | At proficiency | Above proficiency |
|---|---|---|---|
| Core standard | Same target | Same target | Same target |
| Scaffolding | Steps made explicit, worked examples, sentence frames | Moderate supports, some steps prompted | Minimal; students structure their own approach |
| Reading load | Reduced complexity, key vocabulary pre-taught | Grade-level text | Extended or additional sources |
| Practice | Guided, chunked, more repetitions | Standard set | Fewer routine items, more extension and transfer |
| Challenge | Reach the standard with support | Meet the standard independently | Apply the standard in a novel context |
Because the workflow is grounded in the Learning Commons Knowledge Graph — the standards for all 50 states plus the learning progressions beneath each one — the "below" tier isn't just "the same worksheet with easier numbers." Claude knows which prerequisite competencies sit under the standard, so the scaffolds it builds are the right scaffolds: the specific earlier skills a struggling student likely needs shored up. And the "above" tier extends along the progression rather than off into unrelated trivia. The graph is what keeps all three tiers honestly aimed at the same place.
Driving the workflow well
The quality of the tiers depends on the quality of what you tell Claude about the room. A generic "differentiate this" gets you generic tiers. The teachers who get precise, usable materials give Claude the same picture they carry in their heads.
Tell it who's actually in the class. If you know three students are English learners, ask for the below-level materials to pre-teach vocabulary and include the practice in both English and Spanish where that fits. If two students have IEP goals around working memory, say so, and the scaffolds will chunk accordingly. Name the specific student needs, not just the abstract levels — the skill is explicitly built to adapt for particular needs, not only for a below/at/above ladder. And ask for the differentiation plan alongside the materials: the plan tells you which student gets which version and why, which is what turns a stack of worksheets into an instructional decision you can defend.
Differentiation done right means every student works toward the same standard — the scaffolds change, the destination doesn't. A tool that quietly lowers the goal for some students isn't differentiating; it's giving up on them politely.
Once the materials come back, revise them the way you'd revise any draft. You know your students' names and histories; Claude knows the standard and the progression. Push on anything that reads as too easy or too far a reach, and it re-tiers against the same core content. The materials render out to clean, printable documents, so the version you approve is the version you hand out — not a chat window you have to reformat during your prep period.
Where the workflow connects
Differentiation isn't a solo act; it's a node in a loop. It works best downstream of knowing where students are — the class-data workflow lets you hand Claude diagnostics and rosters first, so the tiers map to real students rather than assumed ones. It also pairs with connectors: Diffit is built to create and adapt instructional materials for every student, and its integration with Claude is a natural companion for the manual differentiation moves you'll make between full re-tierings.
If you want the argument for why solving differentiation is such a high-leverage win, One Lesson, Three Readiness Levels makes the case that this single task is the highest-value, most time-expensive thing on a teacher's plate. And the deeper Diffit + Claude workflow shows how connector-driven adaptation fits alongside the built-in skill. For teachers experimenting outside Claude for Teachers, the marketplace education agent skills collection covers adjacent ground.
The equity argument, made practical
There's a reason this workflow matters beyond convenience. Differentiation is where the gap between what we know works and what a teacher's week allows is widest — and that gap falls hardest on the students who most need the below-level scaffolds and the above-level stretch. When writing three versions of every lesson is a fantasy, teachers rationally write one, aimed at the middle, and the tails go underserved. Collapse the time cost and the calculus changes. Every student can get materials pitched at where they are, every day, without their teacher losing another evening to it.
That's the promise the Claude for Teachers series keeps returning to: research-backed practices like differentiation reliably help students, but they're expensive in exactly the currency teachers have least of. Give the time back, and the practice becomes possible. Verification is free through June 30, 2027 for a full year — and if you're going to start with one workflow, differentiation is the one that lets you finally teach to the whole room.
Part of the Claude for Teachers series. Related: One Lesson, Three Readiness Levels: The Differentiation Problem, Solved · Diffit + Claude: Adapting Instructional Materials for Every Student. Browse AI tutoring skills or more builder insights.