Every AI lesson tool makes the same claim
Open any AI planning tool and you will find the same promise: sound pedagogy, research-backed, aligned to how students learn.
Almost none of them prove it. The pedagogy is a line on a landing page, not something you can point to inside the product.
We decided to do the opposite — to make the pedagogy something you can check.
What "good pedagogy" actually means here
The framework NSW schools — Catholic dioceses and the Department of Education alike — are moving toward is explicit instruction. Teach first, then check for understanding, then practise. Clear learning intentions. Worked examples. Checking every student, not just the confident three. Retrieval and review.
It is not a style or an opinion. It is the model with the strongest evidence base in education, and it is what a growing number of school systems now train their teachers to use.
The problem: AI drifts the other way
Left to its own devices, a general-purpose AI writes lessons that drift toward discovery learning — "what do you notice?", "explore in groups", "share your initial thoughts". It reaches for the exploratory mode by default.
For a NSW teacher whose school has committed to explicit instruction, that output is not a starting point. It is a rewrite.
What we did: put the rulebook in the code
Instead of asking the AI nicely to follow explicit instruction and hoping for the best, we wrote the rules down — and built them into the software as a check that runs on every single lesson.
The rules come straight from the two texts these school systems train their teachers on:
- •Anita Archer & Charles Hughes, *Explicit Instruction: Effective and Efficient Teaching*
- •John Hollingsworth & Silvia Ybarra, *Explicit Direct Instruction*
We read both, cover to cover, and turned what they teach into checks a computer can run on a lesson in under a second.
What actually gets checked
Every lesson Smart Syllabus Assistant generates is examined against rules like these before it reaches the teacher:
- •**Does the practice match the objective?** If the lesson says students will *calculate*, the independent practice has to make them calculate — not read about it.
- •**Is it teach-first, not discover-first?** No "what are your initial thoughts" openers.
- •**Is the checking real?** "Do you understand?" is not a check for understanding. A specific question with a knowable right answer is.
- •**Is the modelling done properly?** The teacher works one example, then the student immediately does a matched one — not a vague "let us do this together".
- •**Concept before skill, non-examples to draw the boundary, a reason it matters, an attention cue to open, a real review to close.** The quiet mechanics that separate a strong lesson from a plausible-looking one.
Checked — and rebuilt
This is the part that makes it more than a label. When a lesson breaks the rules that matter most, it does not simply get a warning. It gets rebuilt — automatically, before the teacher ever sees it. The tool corrects itself against the model.
The result is a lesson that already holds up to the framework a NSW school is being trained on, instead of one a teacher has to reshape on a Sunday night.
Why this matters
A generic AI tool skips straight to writing the lesson. It never read the books. It cannot tell a strong lesson from a plausible-looking one, because it was never given a standard to measure against.
Putting that standard into the code is the difference between a tool that *sounds* right and one that is accountable to something. For a school deciding whether an AI tool is safe to put in front of its teachers, that accountability is the whole question.
What is next
This is not a finished project — it is a foundation we keep tightening. Every few weeks the checks get sharper and the lessons get stronger, at no extra cost to the teachers already using it.
If your school is moving toward explicit instruction and wants planning tools that move with it, this was built for exactly that.
Smart Syllabus Assistant — NSW-aligned lesson planning, verified against explicit instruction. See a sample lesson or try it free.