AI Conceptual Understanding Generator

How it works

From concept name to a set of diagnostic probes that reveal what students truly understand.

Enter the concept and grade level

The teacher types the concept they want to probe ('dividing fractions,' 'Newton's second law,' 'supply and demand,' 'the carbon cycle,' 'narrative structure') and specifies the grade level and subject. The AI identifies the core mathematical or conceptual structure underlying the topic, the standard procedural steps students are typically taught, and the conceptual relationships that underlie those steps.

AI generates procedural and conceptual versions of the same task

For each concept, the AI generates two types of questions in parallel: procedural questions that test whether the student can execute the steps correctly, and conceptual probes that test whether the student understands why those steps work. A student who can divide fractions but cannot explain why 'flip and multiply' works has procedural knowledge only. The probe reveals that gap. Based on the research framework of Hiebert and Carpenter on mathematical knowledge.

AI generates why, what-if, and explain questions

Ms. Osei teaches 7th-grade math. She generates a conceptual understanding probe for 'finding the area of a triangle.' The AI's probes can include: a standard procedural problem (calculate the area), a 'Why does this work?' question (why do we divide by 2 in the formula?), a 'What if we changed X?' question (what happens to the area if we double the base but halve the height?), and an 'Explain to a younger student' prompt (how would you explain why the triangle area formula works?). Four questions. Four different types of evidence about what students actually understand.

Use probes for formative assessment, class discussion, or unit review

Conceptual understanding probes are versatile. They work as formative assessment items in exit tickets, as discussion starters at the beginning of a lesson, as unit review questions that push beyond recall, or as interview questions for a diagnostic conversation with a student who is struggling. The draft includes guidance on how to interpret student responses: what a procedural answer, a growing answer, and a deep answer look like.

The procedural trap problem

Most assessment in schools measures procedural accuracy, can the student execute the correct steps? A student can achieve a perfect score on a procedural assessment while having no conceptual understanding of what they are doing. That student will struggle when problems are presented in a new form, when procedures need to be adapted, or when they reach the next level of the subject where conceptual understanding is assumed.

Writing good conceptual probes requires deep pedagogical content knowledge, knowing not just the subject but the specific conceptual structures students need to understand and the specific questions that reveal whether they do. The AI encodes that expertise into a tool any teacher can use.

What the generator produces

Five probe types, each designed to reveal a different dimension of conceptual understanding.

"Why does this work?" questions

These questions ask students to articulate the mathematical or conceptual justification for a procedure or fact. A student who can solve a problem procedurally but cannot explain why the procedure works does not have conceptual understanding. These questions reveal that gap directly. The AI generates why-questions that are appropriate for the grade level, not requiring formal proof, but requiring genuine reasoning beyond "because that is how you do it."

"What if we changed X?" variation tasks

Variation tasks present a modified version of a problem and ask what changes. 'What happens to the volume of a cylinder if we double the radius but keep the height the same?' A student with procedural knowledge recalculates. A student with conceptual understanding can reason about the relationship without calculation. The AI generates variation tasks calibrated to the specific relationships and dependencies in the concept being probed.

Explain-to-another prompts

Research consistently shows that explaining a concept to someone else is one of the most effective ways to consolidate and reveal conceptual understanding. These prompts can ask students to explain a concept as if teaching it, 'How would you explain this to a student in the grade below you?' or 'What would you say to a student who thinks X?' The quality of the explanation reveals the depth of the student's conceptual model.

Always true / sometimes true / never true tasks

These tasks can present a mathematical or conceptual statement and ask whether it is always, sometimes, or never true, and crucially, why. 'When you multiply two numbers, the result is always larger than both of them.' Always, sometimes, or never? This question reveals whether students have a conceptual understanding of multiplication that includes multiplication by fractions and negative numbers, or only a procedural understanding from whole-number multiplication.

Error analysis tasks

Error analysis tasks present a worked solution with a deliberate conceptual error and ask students to identify and correct it. These tasks are powerful because they require students to evaluate mathematical or conceptual reasoning, a higher cognitive demand than solving a problem themselves. The AI can generate error analysis tasks with errors that target common conceptual misunderstandings specific to the topic, not arbitrary arithmetic mistakes.

Response interpretation guides

The draft comes with a guide for the teacher: what a procedural answer, a growing answer, and a deep answer sound like, and what to do next. This turns the probe from a simple assessment into a diagnostic tool, as the teacher sees more than whether a student can follow the steps.

Who uses the conceptual understanding generator

Math teachers at all levels use the generator to create conceptual probes for topics where procedural success often masks conceptual gaps, fractions, proportional reasoning, algebra, and calculus are common examples. The probes become part of regular formative assessment routines.

Science teachers use the generator to move beyond recall-level assessment toward questions that probe causal reasoning, "why does this reaction occur?", "what would happen if we removed this variable?", "explain this result to a classmate who missed the lab."

Instructional coaches and department heads use the generator to calibrate assessment quality across a department, ensuring that assessments across the team probe conceptual understanding, not just procedural recall.

Teachers preparing students for high-stakes exams use the generator because exams like the AP, IB, and SAT regularly include questions that require conceptual understanding, not just procedural execution. Practicing conceptual probes specifically prepares students for those question types.