Snorkl + Claude: Insights on Classes, Assignments, and Progress

Every teacher is drowning in data that never changes a single decision. Assignment scores pile up in the gradebook. Progress reports get generated, printed, filed. Standardized results arrive months after they could have mattered. The problem was never a shortage of information — it's that the information rarely arrives in a form that tells you what to do on Thursday. An insight that doesn't alter a plan is just paperwork with extra steps.

That gap — between knowing how a class is doing and knowing what to teach next — is exactly where Snorkl fits in the Claude for Teachers connector directory. Its job is specific: surface insights about classes, assignments, and student progress to inform instruction. Note the last three words. Not insights for a dashboard. Not insights for a compliance report. Insights to inform instruction. In a product built on the premise that AI should strengthen teaching practice, that framing is the entire point.

Insight is only half the job

It's tempting to treat "insights about student progress" as the finish line. It isn't. Analytics tools have shown teachers colorful summaries of class performance for years, and most of it changed nothing, because a summary is a description of the past, not a prescription for the next lesson. The hard part was always the translation: this cluster of students is stuck on this idea, therefore tomorrow's warm-up should look like this and this small group needs that.

Snorkl's connector is scoped to the surfacing — insights about classes, assignments, and student progress. What makes that valuable inside Claude for Teachers is that the surfacing lands somewhere it can be acted on immediately. The same conversation that shows you the insight is the one that can plan the response, grounded in your standards and curriculum. The insight doesn't get exported, emailed, and forgotten. It becomes the opening move of a planning session.

That placement is the difference between a dashboard and a decision. A dashboard lives in its own tab, and reading it is a task that competes with all your other tasks — so it loses, most weeks, to the more urgent work of tomorrow's lesson. When the insight surfaces inside the tool where tomorrow's lesson is planned, reading it and acting on it are no longer two separate jobs. The friction that usually keeps data-informed instruction aspirational — the export, the interpretation, the manual translation into a plan — is exactly the friction the connector model removes.

From progress to plan, without leaving the conversation

Here is what that continuity buys you. This is the one bulleted list this piece will spend:

• The insight arrives where the planning happens. Snorkl surfaces where a class or a student is; Claude for Teachers is already the place you plan the next lesson. There's no handoff between the tool that knows and the tool that acts — they're the same conversation.
• The response is standards-aware from the first draft. Because CfT is grounded in the Learning Commons Knowledge Graph — academic standards across all fifty states plus the competencies beneath them — an insight about a stuck concept maps to a specific competency and its progression. You don't just learn that students struggled; you get pointed at what comes before and after on the learning path.
• The picture composes with everything else. Progress insights from Snorkl sit naturally alongside a folder of diagnostics, attendance, and notes. That's the same raw material the analyze class data to plan instruction workflow draws on when Claude builds a picture of where each student is.
• The next step can run itself. Once "insight leads to adjusted plan" is a repeatable loop, CfT can schedule it — reviewing progress and adapting tomorrow's plan on a cadence, so the review happens at 4pm while you drive home.

Read those four points together and you see the shape of the whole system: surface, ground, compose, automate. Snorkl owns the first; Claude for Teachers wires it to the rest.

Insight and diagnosis are different tools

It's worth drawing a clean line between Snorkl and its sibling connector, because they answer different questions. A diagnostic question — the job Eedi does — probes a single concept to reveal why a student thinks what they think, in the moment. Snorkl's insights operate at a different altitude: across classes, assignments, and progress over time. One is a microscope on a misconception; the other is a map of the terrain.

You want both, and you want them for different decisions. Use progress insights to decide where to point: which class, which assignment, which trend deserves your attention this week. Use diagnostic questions to decide what to teach once you've pointed there. Confusing the two — trying to make a progress trend do a misconception's job, or vice versa — is how "data-informed instruction" turns into busywork. The connector directory gives you the right instrument for each altitude.

Data that changes no plan is just paperwork. The test of any classroom insight isn't how detailed it is — it's whether it moves next Thursday's lesson.

Insight you control

There's a privacy dimension here that teachers should read carefully, because progress insights necessarily touch student information. Claude for Teachers is unambiguous on this: you control what's shared, and nothing shared is used for model training. Student information is protected under a K-12 Data Processing Addendum written to comply with FERPA, and the product runs on its own teacher terms built for K-12 privacy.

That matters more for a progress-insight connector than almost anywhere else, because progress data is inherently about identifiable kids. The design answer isn't "trust us" — it's structural: teacher-only access consistent with Claude's 18-and-over policy, contractual FERPA protection for student data, and no training on what you share. When you're deciding how much class context to bring into a planning session, those guarantees are what make it a responsible decision rather than a leap of faith. If you want a practical framing for what to share and what to hold back, the series covers a data-hygiene approach for exactly this.

The loop this closes

Step back and the value is a loop that most schools have never been able to close. Assess or observe. See where students are. Adjust instruction. Repeat. Research-backed practice — mastery-based pacing, small-group instruction, differentiation — depends on running that loop tightly and often. What made it infeasible wasn't ignorance of the loop; it was the labor of turning raw progress data into an actual instructional adjustment, every cycle, for every group.

Snorkl surfaces the insight. Claude for Teachers, grounded in real standards, turns it into a plan, composes it with your other class data, and can even run the review on a schedule. That's the loop closing — not because a teacher works harder, but because the translation from what happened to what's next stops being an evening's manual labor. If you're building out the assessment side of that loop, the marketplace's exam blueprint skill and the broader education agent skills pair well with a progress-insight connector.

Want the map of how these connectors compose? Start with the series hub, or browse more builder insights.

Part of the Claude for Teachers series. Related: Eedi + Claude: Diagnostic Questions That Reveal Student Thinking · Hand Claude a Folder: Analyzing Class Data to Plan Instruction. Browse AI tutoring skills or more builder insights.