How It Works
Learning Outcomes
& Progress Tracking
Define what mastery looks like. Let AI track each student's journey toward it — turn by turn, in real time, grounded in educational science.
From Snapshot to Dialogue.
Traditional assessment offers a single snapshot of proficiency at a fixed point in time. A well-designed AI chatbot does something fundamentally different: it facilitates a live argument between the learner and the system, where the conversation itself becomes the primary instrument for assessing complex cognitive outcomes. Secondary AI is built on this principle — tracking is not bolted on, it is the architecture.
"Assessment embedded in every exchange, not reserved for the end."
— The design principle behind Secondary AI's outcome tracking
The Foundation
Start
with the end.
Secondary AI is built around Understanding by Design (UbD) — the gold standard for curriculum planning for over two decades. You begin by defining the specific learning outcomes you want students to achieve before a single conversation begins.
The AI then monitors every message, continuously assessing whether each outcome has been addressed, and steering students toward goals they haven't yet reached — not just the ones they choose to ask about.
Example
outcomes.
Biology
- Explain the role of ATP in cellular respiration
- Distinguish between aerobic and anaerobic processes
- Connect mitochondrial function to overall cell health
History
- Analyze causes of the First World War using primary sources
- Evaluate the effectiveness of the Treaty of Versailles
The Science Behind It
Six pedagogical frameworks.
One coherent system.
Secondary AI's outcome tracking is not a feature — it is a convergence of the most rigorous frameworks in educational measurement and learning science. Each framework plays a distinct role in transforming dialogue into defensible evidence of mastery.
Social Constructivism & ZPD
The AI acts as a 'More Knowledgeable Other', keeping students in their Zone of Proximal Development — challenged but supported. Errors and misconceptions aren't failures; they're the most active sites of learning, captured as data.
Socratic Method & Bloom's Taxonomy
Rather than testing recall, the AI drives hierarchical questioning — climbing from remember and understand through to analyze, evaluate, and create. Each student turn is a window into their cognitive depth.
Evidence-Centered Design
Every conversational turn is treated as telemetry — evidence of latent competency. The system links observable student behaviors directly to the learning outcomes you defined, producing defensible, longitudinal mastery claims.
Adaptive Measurement
Drawing on Item Response Theory, the AI calibrates question difficulty to each student's demonstrated ability — maximizing precision at their current level, avoiding boredom and frustration simultaneously.
Continuous Formative Assessment
Assessment is embedded in every exchange, not reserved for the end. Immediate, individualized, process-focused feedback acts as feedforward — guiding the next step rather than judging the last.
Cognitive Load Management
By providing targeted real-time prompts and scaffolding, the AI minimizes extraneous cognitive load — freeing students to focus on meaningful problem-solving rather than guessing where they went wrong.
Assessment Reimagined
Every conversation turn is evidence.
Traditional Testing
Response Format
Static, discrete items — MCQs, fill-in-the-blank.
Data Granularity
Outcome-focused: correct or incorrect.
Timing
Terminal — at the end of a unit or course.
Feedback Loop
Delayed, focusing only on what was wrong.
Integrity Model
Policing and detection of cheating.
AI Formative Assessment
Response Format
Dynamic, free-text dialogue across non-linear turns.
Data Granularity
Process-focused: turn-by-turn reasoning and path selection.
Timing
Ongoing — embedded in every single interaction.
Feedback Loop
Immediate, offering hints and adaptive next-step guidance.
Integrity Model
Process documentation — resistant to simple cheating.
What Students & Teachers See
Real-time progress.
For every student.
As students chat, the AI continuously assesses each message against your defined outcomes, updating a live progress indicator the student can see. This transparency is intentional — research shows that making learning goals visible significantly increases metacognitive awareness and self-directed learning.
The AI will not abandon an outcome until it is confident the student has genuinely demonstrated mastery. Surface-level or guessed answers are probed further — building true understanding rather than compliance.
95% confidence
60% confidence
0% confidence
Teacher Dashboard
Insights that drive
your next lesson.
Advanced dashboards provide intuitive visualisations of individual progress trajectories, class-wide learning analytics, and curriculum effectiveness metrics — enabling early identification of students at risk and timely interventions previously unattainable at scale.
Live outcome progress per student, per session
See exactly where every student is on every outcome at any moment — not after the test, but during the conversation.
Class-wide outcome gap analysis
Identify the outcomes that most of the class hasn't yet mastered. These surface automatically as your highest-priority reteaching targets.
AI-generated reteaching suggestions
The system flags patterns of misconception across the class and suggests targeted instructional adjustments for your next lesson.
Rubric-aligned grading integration
Attach your defined outcomes directly to rubric criteria for perfectly aligned summative grading — no manual cross-referencing needed.
Competency-Based Education ready
For CBE schools, the AI tracks mastery rather than time-on-task — enabling genuine competency-gated progression built into every session.
An Important Caveat
AI tracks. You judge.
Research consistently shows that AI-assisted assessments increase the speed and depth of formative data — but do not replace the need for human-mediated conceptual understanding. Comparative studies also reveal that AI tends to assign higher confidence scores than human instructors in high-stakes scenarios. All outcome tracking in Secondary AI is explicitly designed as formative support, and human teacher review and sign-off is required before any grade is recorded. The future of educational assessment lies not in the replacement of human intelligence, but in its strategic augmentation.
Track Every Student's Mastery Journey
Set up your learning outcomes and let the AI do the tracking — you focus on the teaching.
Get Started Today