01
Diagnose
Start from what the cohort already shows you: mastery per skill and per depth of knowledge, drawn from the practice your students are already doing. No separate placement test to administer.
Available by request
Learn adaptively. Teach adaptively.
Your students already get an adaptive path. Your group time, the most expensive minutes in education, still runs on guesswork. Adaptive Teaching reads your cohort's knowledge map and builds a Mastery Spiral: short, structured lessons in dependency order, targeting the foundational gaps that block everything above them. You stay in charge. It does the homework.
One loop of the Spiral, repeated until the foundations hold.
A Spiral is a series of units. Each unit is a run of lessons that ends in a checkpoint. Every pass through the cycle revisits skills at greater depth, from recall to application to strategic thinking.
02
Plan
The engine selects the highest-impact teachable skills and orders them by the prerequisite graph, foundations first. If a cohort is not ready for a skill, the plan drills down to the blocking foundation instead of teaching over their heads.
03
Teach
A unit of short, structured lessons, 30 minutes by default. Each arrives with its objective, a worked example, a practice set, and the misconceptions this specific cohort keeps hitting.
04
Checkpoint
Every unit closes with a cumulative, interleaved review built from real practice data, read alongside everything else your students did that cycle. A single quiz never decides a student's path.
05
Synthesize and re-plan
Results reshape the next unit: what to advance, what to reteach with a changed approach, who needs a different group, and what routes to individual practice. You approve the change before anything moves.
What you get
Every lesson arrives prepped, with its reasoning attached.
Prep a full session in minutes instead of an hour, and know exactly why the platform put this skill in front of this group today.
Objective and depth of knowledge
What students should be able to do by the end, and at what cognitive depth: recall, application, or strategic thinking. Stated in student language so you can put it on the board.
Why this, why now
Every lesson shows its reasoning in plain numbers. For example: 9 of 12 students below mastery, 11 of 12 ready to learn it, and it unlocks 5 downstream skills. No black box.
A worked example and a practice set
One example worked step by step, plus a fixed set of questions drawn from the same bank your students already practice on. It is a teaching aid to draw from, never a quota to finish.
The misconceptions your cohort actually has
Educator-only notes naming the most-missed questions and which wrong answers students are picking, so you can address the specific error rather than the generic one.
Exit ticket and follow-up work
A short check to close the session, plus pre-work for students who are not ready yet and consolidation for the ones who wobbled. One click each.
Materials, vetted and attributed
Videos, articles, and interactives with their duration and depth level marked, linked or embedded with attribution. Never re-hosted.
Every minute planned, honestly.
The default session is 30 minutes, which suits one to three students and a tight scope. Larger groups and device-heavy sessions get a recommendation of 45 to 60 instead. Question counts shrink before the pace does.
- 5 min Warm-up retrieval Two or three questions on previously taught skills, always including the last lesson's target.
- 1 min Framing State the objective in student language and keep it visible for the rest of the session.
- 12 min Mini-lesson Explicit instruction with one worked example, making the thinking visible rather than just the answer.
- 9 min Guided practice Students get at-bats on the practice set while you watch the work, with answers visible to you only.
- 3 min Exit ticket A three or four question check sent to each present student, with a live completion ticker.
Sessions work offline too. Present or print the plan, teach it, and mark it complete. Devices add exit tickets and live tracking, but nothing requires them.
The rules the system will not break.
The engine plans. AI only narrates. You approve.
A deterministic, auditable engine decides what gets taught and in what order. AI is limited to drafting the prose of a lesson plan, it is clearly labeled as a draft, and nothing reaches a classroom until an educator approves it. AI never decides curriculum.
You name your groups. We only suggest.
Group names belong to you. Each new group arrives with a neutral two-word name, for example "Brave Falcons", and you rename it inline to whatever your practice or school already uses. The skill being worked on sits underneath as a subtitle, and your name follows the students through a re-plan. Underneath, there is no rank or ability field, so grouping stays a readiness split and never a hierarchy. Groups re-form from fresh data at checkpoints, never mid-unit, and stability is treated as a pedagogical asset rather than something to churn.
Group time is the scarcest resource, so it is spent deliberately.
Group minutes go only to gaps that are shared, foundational, and teachable. Anything idiosyncratic routes to that student's own practice instead. When a group lesson pushes a skill past mastery, the student's personal plan stops serving it automatically.
A missed skill is retaught differently, not louder.
A skill that does not land carries to the front of the next unit, always before anything that depends on it. The reteach must change something: a different example, a different method, a scaffold, a smaller group. After a bounded number of attempts it routes out of group time rather than consuming more of it.
Predictions are made up front and measured after.
Every plan states its expected outcomes before you teach, and every checkpoint measures against them. You see expected versus actual per skill and per student, and the planner calibrates itself as evidence accumulates.
Engagement is evidence, not decoration.
Speed on right versus wrong answers, rapid guessing, hesitation, and answer changes are all tracked, because outcome data without engagement context misattributes failure. Slow and wrong is a different problem from fast and wrong.
Research foundations
The defaults come from the evidence, not from a hunch.
High-dosage tutoring
The evidence base points to three or more sessions a week of 30 to 60 minutes, sustained over about ten weeks, in small groups. Those are the defaults the planner starts from, rather than arbitrary session lengths.
Bruner's spiral curriculum
Skills are revisited at increasing depth rather than covered once. A skill taught at recall level returns later at application, then at strategic thinking, which is what makes the Spiral a spiral.
Interleaved and cumulative practice
Mixing skills within a review beats blocking them, even though it feels harder at the time, and the advantage shows up on delayed tests. Checkpoints are cumulative and interleaved by design for exactly this reason.
Multidimensional readiness over tracking
Students with very different total scores often show equivalent readiness for a specific skill. Grouping runs on per-skill readiness profiles rather than a single overall score, which is why a group is only ever defined by the skill it is working on.
For the mastery model underneath all of this, read our level-based learning guide, or the full Adaptive Teaching guide.
Bring it to your cohort.
Adaptive Teaching is live and rolling out gradually, so access is granted by request. Tell us about your students and we will get you set up.
Currently available to tutors, schools, and programs by request.