AI lesson planning from your adopted curriculum, with every activity cited to a page and a state standard
Every lesson drafted from the adopted curriculum, every activity cited to a standard.
See one case, screen by screen ↓AI Lesson Planning is an AI lesson planning tool that drafts K-12 lessons from the district's board-adopted curriculum, cited to state standards, for teachers. Seven agents draft the week from the adopted teacher's edition and the pacing guide, cite a page and a standard on every activity, check the teacher's edits against the approved variations, add supports as counts, and replan the unit when a lesson is lost. The teacher decides and publishes.
AI in education: a buyer's guide to agentic AI for colleges, universities and school districts →
Updated
AI lesson planning is already part of the week: about a quarter of ELA, math and science teachers used AI for instructional planning or teaching in 2023–24, and most of them used it only for planning. But a generic AI lesson plan generator drafts from the open web, not from the high-quality instructional materials the board voted to adopt and the district trained teachers on. Lessons drift from the adopted sequence, and a coach finds out at a classroom visit, weeks later. Student names end up in prompts.
The calendar does the rest. A fire drill, an assembly or a lab that runs a second day costs a lesson, and the teacher rebuilds the unit against the pacing guide by hand, usually by cutting the end of it, which is often the engineering design work. Accommodations from IEPs and Section 504 plans reach teachers as at-a-glance sheets, and sentence frames in a student's home language depend on the teacher's spare time. AI Lesson Planning puts seven agents on the teacher's week, drafting only from the adopted curriculum, while the teacher decides what reaches students.
Estimated split of one secondary teacher's weekly planning hours, by hand and with the agents; about three hours a week back, in line with the target. Our arithmetic, kept modest.
A Planner drafts from the adopted unit, a Fidelity agent checks every edit, a Catch-up agent replans; the teacher publishes.
Elena Ortiz-Lang teaches Grade 7 science to 148 students in five sections at Linden Grove Middle School, Marlow Bay Unified. Unit 3, Energy transfer, runs 15 school days to a unit test on Monday, November 2. On Thursday, October 22, a fire drill takes the period. Here is what happens next, screen by screen, in the working solution.
Elena opens her week as five day columns. The Planner agent drafted five lessons with 61 citations: each block shows its standard codes, its teacher's edition pages and a timed list of activities. Support markers read "EL 23 · ES VI" and "IEP · 504", counts and types, never names.
Day 8 of 15 · 2 days behind · unit test Mon, Nov 2
At 10:58 a.m. Elena marks Lesson 7 not taught. Four minutes later the Catch-up agent has read what happened, the pacing guide and the school calendar, and draws its plan on next week as dashed blocks, each with what it replaced and why. It folds Lessons 9 and 10 because the teacher's edition allows it when short on time (TE p. 146).
Still one day short: Lesson 14 has no day.
The agent offers three ways to recover the last day, with what each recovers: swap the two-day cup lab for a teacher demonstration (+1 day), shorten the Lesson 14 review to a warm-up (+½ day), or ask the principal, Dr. Ana Whitcombe, to move the unit test to Tuesday (+1 day, needs approval). Elena picks the demonstration.
The Fidelity agent checks the swap against the register of approved variations: within the approved variation, TE p. 157, which allows a teacher demonstration when time is short. It names the practice that moves: MS-PS3-3 keeps "design" through the paper design and "test" through the demonstration, and "construct" moves to the optional build at home. The Standards agent re-runs coverage: all 8 unit standards are taught and assessed by Mon, Nov 2.
This plan is back on pace by Wed, Oct 28.
The Access agent drafted Spanish frames for 19 English learners at 0.97 confidence and a Vietnamese frame for 4 newcomers at 0.88. Below the 0.90 threshold, the frame goes to a bilingual specialist, not the teacher; Thanh approves it at 11:38 a.m. Accommodations appear as types and counts, from 18 IEPs and 7 Section 504 plans.
Elena accepts the plan. The pacing bar and the week update at once: Lessons 9–10 on Tuesday, the paper design and demonstration on Wednesday, Lesson 13 on Thursday, the review on Friday. The change goes to the unit log her coach reads.
Back on pace by Wed, Oct 28 · unit test Mon, Nov 2 held
Lessons 9–10, Heat on the move and the cup challenge, is a 50-minute period of timed activities: a 5-minute warm-up, a 12-minute mini-lesson on three ways heat moves, two stations instead of three, and the cup challenge brief. It carries 14 citations, and all resolve: clicking TE p. 142 opens the page with the passage highlighted.
Elena adds an outside video link to "Two stations, not three". The Fidelity agent re-checks the edit: the link is not in the adopted materials, and the unit has its own convection video (TE p. 161). She can use the unit video, or keep her edit with a reason for her coach. She uses the unit video.
Nothing reaches students until she publishes. She picks the five sections and a release time of Tuesday, 7:30 a.m., before period 1. The Publish agent checks the content for student names and IDs and finds none; the learning management system assigns the versions with read-aloud, checklists and extended time from the roster.
Darius sees the change in his queue: Lessons 9–10 folded, the cup lab swapped for a demonstration, 14 citations, each one opening its passage. He reviews the citations; coaches comment and review, but they do not edit or publish a teacher's lesson.
The same unit, seen by the five people who carry it: what their week looked like, and what it looks like with AI lesson planning on the adopted curriculum.
A Planner drafts from the adopted unit, a Fidelity agent checks every edit, a Catch-up agent replans; the teacher publishes.
Drafts a week or a lesson from the adopted unit, the pacing guide and the class's support counts, and cites a page and a standard for every activity.
Compares every draft and every teacher edit with the board-approved materials and their approved variations, and names the standard practice a change affects.
Applies accommodation types and English learner supports at the lesson level, and drafts home-language sentence frames beside the English.
Replans the rest of the unit when a lesson is lost, keeps the unit test date, re-runs coverage and offers options with what each recovers.
Maps every activity to state standards and checks the unit's coverage, taught and assessed, after every change.
Packages slides, handouts, frames and the exit ticket, checks that no student name or ID is in the content, and publishes when the teacher says.
Builds board, state reviewer and privacy packs from the audit trail, pseudonymizes student IDs and seals each pack with a hash.
Publish, review, approve sharing. The agents propose; a named person decides.
The assistant answers from the unit's own sources, cites them, and opens the screen that acts on the answer.
Why did the Catch-up agent fold Lessons 9 and 10?
The teacher's edition marks Lesson 9 to be combined with Lesson 10 when time is short, and the pacing guide allows it. The fold recovers one day and keeps MS-PS3-3, MS-PS3-4 and MS-ETS1-1 taught and assessed.
Which Unit 3 standards are at risk?
Before the catch-up plan, MS-PS3-5 was at risk: Lesson 14 had no day before the test. With the plan accepted, all 8 standards are taught and assessed by Mon, Nov 2.
What supports does Lessons 9–10 carry?
Five accommodation types, as counts: read-aloud (9), chunked steps (11), graphic organizer (14), extended time (8) and speech-to-text (2, new Oct 20). Frames in Spanish for 19 and Vietnamese for 4 newcomers. No student is named; plans stay in the case system.
Show every change to Unit 3 this week
Since Monday: the Lesson 5 lab ran a second day; Lesson 7 was lost to the fire drill; the Catch-up agent proposed a plan; Thanh Nguyen-Brooks approved the Vietnamese frame. The audit trail opens filtered to Unit 3.
13 screens from the working solution, on its sample data. Pick one to see it large.
Thresholds, review times and retention are settings, versioned and on the audit trail.
| Setting | Default | Choose from |
|---|---|---|
| Citation confidence | 0.85 | Below it, the activity is not drafted · Planner and Standards agents |
| Translation confidence | 0.90 | Below it, an English learner specialist reviews the frame |
| Coach citation review | 3 school days | Escalates to the principal, then the director |
| Bilingual frame review | 1 school day | Escalates to the EL director; English-only frame until approved |
| District share approvals | 5 school days | Reminder at 3 days; escalates to the director |
| Fidelity flag decision | 2 school days | Escalates to the content coordinator |
| Keep drafts and conversations | 3 years | Audit trail kept separately, minimum 7 years |
| Prompt, model or threshold change | Administrator and director together | Versioned; every override carries a reason code |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
Sources: RAND, uneven adoption of AI tools among U.S. teachers and principals (February 2025; about a quarter of ELA, math and science teachers used AI for planning or teaching in 2023–24, and 64% of those used it only for planning) · Gallup and Walton Family Foundation, Teaching for Tomorrow (June 2025; 5.9 hours a week, self-reported by weekly AI users, across tasks) · Standard codes are public Next Generation Science Standards codes · "target" = design goal, measured in the live solution · "demo" = seen in the working solution, on its sample data · "typical" = published figures · "estimated" = our arithmetic, kept modest. Marlow Bay Unified School District, its schools and the people named on this page are fictional: characters and sample data in the working solution.
AI Lesson Planning is an AI lesson planning tool that drafts K-12 lessons from the district's board-adopted curriculum, cited to state standards, for teachers. Seven agents draft the week from the adopted teacher's edition and the pacing guide, cite a page and a standard on every activity, check the teacher's edits against the approved variations, add supports as counts, and replan the unit when a lesson is lost. The teacher decides and publishes.
The agents draft only from a source library the Director of Curriculum owns: the adopted teacher's editions, pacing guides, state standards and the board's approved variations. Every activity cites its page and standard, and clicking the citation opens the passage. The Fidelity agent re-checks every draft and every teacher edit, and flags anything outside the adopted materials with the approved alternative.
A generator makes one document from a prompt. Here the teacher's week is the canvas: lessons are paced across the unit, standards coverage is re-run after every change, supports come from the roster as counts, and the teacher's coach sees each change in the unit log. Nothing reaches students until the teacher publishes.
When a lesson is marked not taught, the Catch-up agent replans the rest of the unit and draws its plan on next week, keeping the unit test date. If the plan is still short, it offers options with what each recovers, such as swapping a two-day lab for a demonstration the teacher's edition allows. The teacher chooses; a test date move needs the principal.
The agents use roster counts, English learner proficiency and home language, and accommodation types, never student names, IDs, grades, discipline or IEP narratives, and no model is trained on district data. Accommodations appear as types and counts, and student-level plans stay in the special education case system. Home-language frames below 0.90 confidence go to a bilingual specialist before students see them.
Yes. Agents propose and cite; the teacher accepts, edits, turns a plan down with a reason, or asks back, and only the teacher of record publishes. Coaches review and comment but do not edit or publish a teacher's lesson, and a district exemplar needs the coach and the director, never the author.
The learning management system for publishing, the curriculum library, the student information system for rosters, sections and the calendar, the special education case system for accommodation types only, the state standards database and the district's single sign-on.
The Agentic Solution Engine builds and deploys it from your requirements and documents: your adopted units, pacing guides, approved variations, roles and systems. It goes live once every quality gate has passed.
Plans drafted in minutes, and no student judged on sessions that never happened.
1-min demoAI Essay FeedbackStudents get feedback in days, not weeks, with every comment signed by their teacher.
1-min demoLiteracy & Dyslexia ScreeningEvery flagged reader's family told in its own language, inside the state deadline.
1-min demoWe'll run AI Lesson Planning on one of your adopted units and its pacing guide.