AI suggests. Teachers decide.
Co-Teacher is the AI assistant for teachers. It drafts the comments, ranks the priorities, and surfaces the alerts — but every output is reviewable, editable, and overridable. Teachers ship; the AI assists.
The teacher is the load-bearing role.
Most ed-tech treats the teacher as the bottleneck — the slow human between the AI and the student. We treat the teacher as the load-bearing role: the one who sees the kid as a kid, not a data point.
Co-Teacher exists to remove the busywork — the report-card paragraph, the parent-email draft, the “who needs help this week” triage — so the teacher gets that time back for the part of the job a model can’t do.
Every AI output is a draft. The teacher is the publisher. That principle is wired into the product, not the marketing.
The Friday morning workflow.
Twelve minutes between opening the laptop and having Monday planned, two students intervened with, and a parent conference booked. Here’s how it actually goes.
- Friday · 07:42
You open CurioPilot. The Co-Teacher digest is waiting: six action cards ranked by impact for the week ahead.
- Friday · 07:44
Card 1 — Critical:“Reteach equivalent fractions to Section 5A. Three students at risk.” You click. A 5-minute reteach mini-lesson is drafted, anchored to the misconception (“treats numerator and denominator independently”).
- Friday · 07:48
You edit two sentences in the lesson — your phrasing for the visual model. You assign it. The audit log records who drafted, who edited, when.
- Friday · 07:52
Card 2 — Warning:“Maya’s risk score crossed 70. Suggest a 1:1.” You click. The conference brief is ready: strengths, gaps, recent misconceptions, recommended next steps.
- Friday · 07:54
You print the brief. You schedule the conference for Tuesday. The parent gets a calendar invite drafted by AI; you review the message, approve, send.
- Friday · 07:54
Time spent: 12 minutes. Three more cards triaged. You decide which to tackle now and which to do during your prep period.
What Co-Teacher does for you.
Co-Teacher Friday digest
Six Action Cards per week, ranked by what'll move the needle most. Each card cites the data: how many students at risk, on which topic, with a recommended next step. Read it Friday afternoon, plan Monday in fifteen minutes.
Real-time micro-alerts
When a student spikes in risk mid-day — failed three quizzes in a row, dropped two Bloom's levels, abandoned an essay halfway — the teacher gets a toast notification. Dismissible, escalatable, or routed to an intervention. Nothing is broadcast to the whole class.
AI-drafted comments
Report-card comments, intervention notes, parent-message drafts. AI generates the first pass grounded in the student's actual data; the teacher edits and approves. Nothing leaves the platform without teacher review — every draft is logged in TraceLayer.
Action Cards (admin view)
School admins see the same Action Card model triaged across the whole school. Severity-tagged. Routed to the right teacher with the right context. The leadership team works the same surface as the classroom — no parallel reporting layer to maintain.
Grade override + rubric approval
Every AI-generated grade or rubric is editable. Teachers can override with reasoning that's logged in TraceLayer alongside the original AI decision. The override path is one click — not a buried flow that nudges teachers toward accepting the default.
How teachers stay in control.
Can a teacher dismiss the Friday digest?
Yes. The digest is a tool, not a directive. A teacher can dismiss any Action Card with a one-line reason (or none), and Co-Teacher learns from that — cards that get dismissed for the same reason three weeks running stop showing up.
Who decides the final grade — the AI or the teacher?
The teacher, always. AI grades are drafts. The teacher's score is the score that goes to the student and parent. The original AI grade is preserved in TraceLayer for audit, but it's not visible to the student or parent unless the teacher chooses to share it.
How does the real-time alert avoid being noisy?
Alerts are gated by an importance threshold tuned per class. A single bad quiz doesn't trigger an alert — a pattern does (three failures in a row, a Bloom's-level regression, an abandoned high-stakes task). Teachers can adjust the threshold per student.
Does Co-Teacher replace the teacher's professional judgement?
No. Co-Teacher surfaces patterns the teacher might miss in a 30-student class — and drafts the boilerplate. The judgement, the relationship, the override authority all stay with the teacher. The platform is built around that: every AI output is reviewable, editable, and overridable.
See Co-Teacher with your own data.
30-minute walkthrough with you, your IT, and your DPO. We’ll run a sample Friday digest against an anonymised class and show the override + audit flow end-to-end.