◆ The Platform

How Si Math AI Works

Si Math AI is a comprehensive learning platform for SAT, ACT, and EST Mathematics that combines educational expertise, AI technology, personalized learning, analytics, and human support to help students improve their understanding and performance.

There are eight systems inside the platform. They are not eight features on a list — they are one loop. Every question a student asks produces evidence, the evidence produces a diagnosis, the diagnosis produces a plan, the plan produces practice, and the practice produces new evidence. That circuit is what separates a learning platform from an answer engine.

Underneath all eight sits the same body of educational knowledge: the teaching methods, mistake patterns, exam strategies and skill sequencing authored and reviewed by experienced educators and SAT, ACT and EST mathematics specialists. Zero delivers that knowledge. Zero does not invent it.

All eight are built for one field. Si Math AI is an educational platform specialized in American Diploma Mathematics, with deep educational expertise in SAT Math, ACT Math, and EST Math — the mathematics sections of those three examinations, and nothing else. That boundary is what makes the depth below possible; see exam coverage.

Si Math AI is not just an AI chatbot. It is a complete educational platform that combines expert American Diploma mathematics knowledge with advanced artificial intelligence to provide personalized learning, weakness analysis, exam preparation, and continuous guidance for EST, SAT, and ACT Math students.

◆ The Loop

From your first question to mastery

Every question moves through the same five steps. Step three is the one general AI tools do not have, and it is the reason the other four compound instead of repeating.

1. Ask

   Type the question, paste it, or photograph it straight from a prep book or worksheet. English, Arabic or Franco — whichever you think in.

2. Understand

   Zero works through it step by step, offers a second approach when the first does not land, and explains why the wrong answer choices are wrong — because knowing why a distractor is tempting is what stops you picking it next time.

3. Diagnose

   The attempt is recorded against a specific skill in the platform's taxonomy. The Weakness Analyzer re-ranks your weak skills by how much each is costing your score. This happens quietly, on every interaction.

4. Focus

   Focus Practice builds a drill set from that ranking — only the skills actually losing you points, in the order that recovers the most marks per hour of study.

5. Master

   Mastery for the skill rises, the predicted test-day score updates, and the weakness disappears from your dashboard once later evidence confirms the fix held. Then the loop runs again on whatever is now the top weakness.

---

The complete learning cycle

Those five steps are the summary. Below is the same loop in full — every system a single question passes through, in order. This is the difference between a chat window and a learning platform: a question does not stop at the answer, it travels.

1. 📝

   START Student asks a question

   Typed, pasted, or photographed straight from a prep book or worksheet — in English, Arabic or Franco.

2. 🐉

   01 Zero AI Tutor

   Zero works through the problem step by step, offers a second approach if the first does not land, and explains why the wrong answer choices are wrong — delivering the teaching method a specialist authored, not one it invented.

3. 🔬

   02 Question Analysis

   The attempt is resolved to one specific skill in the 33-skill taxonomy and written as a diagnostic signal. This is the quiet step that makes everything below it possible.

4. 🔍

   03 Weakness Analyzer

   The new signal joins every previous one, and weak skills are re-ranked by how much each is costing the student's score, each carrying a severity band.

5. 🧠

   04 Learning Memory

   The question, the explanation and the mistake are stored permanently and searchably. Nothing is discarded at the end of a session, which is why the loop compounds instead of restarting.

6. 🎯

   05 Focus Practice

   A drill set is generated from the ranking — only the skills currently losing marks, in the order that recovers the most per hour. Mastered skills are skipped.

7. 📋

   06 Mock Exams

   Full-length, correctly timed SAT, ACT or EST mocks test whether the fix survives exam conditions — evidence gathered under real exam timing.

8. 📈

   07 Progress Tracking

   Mastery per skill, trend lines and the predicted test-day score all update — making improvement visible, and stagnation visible early enough to act on.

9. 🤝

   WHEN NEEDED Human Support

   Educators and exam specialists review how students actually perform and revise the material; people handle accounts and upgrades. This step is conditional in any single cycle — but it is continuous behind the platform.

10. 🏆

    RESULT Better Performance

    The weakness clears once later evidence confirms the fix held, the predicted score climbs, and the skill leaves the dashboard.

↻  The cycle repeats — each pass starts from a sharper diagnosis than the last

◆ System 01

Zero — the AI Mentor

Zero is the tutoring layer: available at any hour, patient across any number of repetitions, and speaking the language you reason in.

Zero is a fictional AI mentor character — a dragon guide, not a real person, and never presented as one. What matters about Zero is not the character but what sits behind it: Zero delivers explanation methods, worked approaches, mistake diagnoses and exam strategies that human educators and exam specialists created and continuously review. Zero is the interface to that expertise, and the reason it can reach one student at a time at three in the morning.

◆ System 02

Weakness Analyzer

The system that replaces guessing. It tells you which skills are costing you marks, ranked by impact, with the evidence behind each conclusion.

Most students cannot accurately name their own weaknesses. The topics that feel hardest are often not the ones losing the most marks, and the errors that actually cost points — a sign dropped under time pressure, a word problem mistranslated into an equation — are invisible from the inside. The Weakness Analyzer solves this the way an experienced teacher would if they could review every attempt you had ever made.

How the diagnosis is produced

• Every attempt becomes a signal. Questions asked of Zero, focus drills and mock exam items all write a diagnostic signal recorded against a specific skill — not a vague topic label.
• Signals resolve to a canonical skill. The platform maintains a fixed taxonomy so that "quadratics", "quadratic equations" and "solving by factoring" cannot fragment into three unrelated records. Every signal is mapped to one permanent skill identifier before it is stored.
• Weaknesses are ranked by score impact. Not alphabetically and not by how wrong you were — by how much the skill is worth on your exam combined with how reliably you are losing it. A frequently-tested skill you fail half the time outranks a rare skill you always fail.
• Each weakness carries a severity band. A four-level band makes the ranking readable at a glance, so you know what is urgent versus what is merely imperfect.
• The diagnosis decays and re-tests. A weakness is not cleared because you got one question right afterwards. It clears when later evidence confirms the fix held.

The taxonomy underneath

Diagnosis is only as good as the vocabulary it is expressed in. Si Math AI is organised around a specialist-defined taxonomy of 5 topic domains and 33 tracked skills:

Skill identifiers are permanent database keys. A display name can be renamed without breaking a student's history, and every stored record carries the taxonomy version that produced it — so a curriculum revision never silently rewrites what an old diagnosis meant.

◆ System 03

Focus Practice

A drill list built from your actual weaknesses — so that every minute of practice is spent on something that moves your score.

A diagnosis with no action attached is just a more precise way of feeling bad about your preparation. Focus Practice is the step that converts the Weakness Analyzer's ranking into work: it generates a targeted set of practice questions aimed at the specific skills currently costing you points, in the order that recovers the most marks per hour invested.

• Generated from live diagnosis, not a fixed syllabus. Two students on the same day, both eight weeks from the SAT, get different drill sets — because they have different weaknesses.
• It skips what you have mastered. Practising a skill already at high mastery is the most comfortable and least useful thing a student can do. Focus Practice does not offer it.
• It re-ranks as you work. Clearing a weakness promotes the next one. The list is a live queue, not a printout.
• Results feed straight back in. Every drill attempt writes a new signal, so the practice that fixes a weakness is also the practice that proves it was fixed.
• Progress is rewarded. Completed sessions earn XP toward the rank ladder, which exists because sustained daily practice is the hardest part of preparation and deserves visible acknowledgement.

◆ System 04

Mock Exams

Full-length, correctly timed SAT, ACT and EST mathematics mocks. You sit the paper under real timing, then record your result and the mistakes you made.

Practising questions and sitting an exam are different skills. Under real timing, students who know the mathematics perfectly well still lose marks to pacing, fatigue, question triage and the decision of when to abandon a question and move on. A mock exam is the only way to measure that, and the only way to practise it.

What a mock produces

• Real timing and real format for each of the three exams. Pacing is the clearest reason that matters: the ACT's per-question budget is a fundamentally different problem from the SAT's or the EST's, and rehearsing one exam at another's tempo trains the wrong habit.
• A result you record yourself. The platform does not convert a raw count into a scaled score and does not split the sitting by section: you enter the score your own result shows, along with how many questions you answered correctly, wrongly and left blank.
• Every mistake feeds the Weakness Analyzer. A mock exam is evidence gathered under real exam timing — this is where a wrong answer under time pressure separates itself from a wrong answer caused by not knowing the content.
• Mistakes are logged for review. The questions you got wrong remain available afterwards, so the mock becomes study material rather than a score you glance at once.

Exam coverage

Si Math AI is an educational platform specialized in American Diploma Mathematics, with deep educational expertise in SAT Math, ACT Math, and EST Math. Three examinations, one field, and the mathematics sections only — Si Math AI does not cover English, Reading, Science, essay writing, admissions consulting or other school subjects. These are not exams the platform happens to support; they are what it is for. See what we specialize in for the boundary in full.

◆ System 05

Smart Progress Tracking

Mastery per skill, trend lines over time, and a predicted test-day score that updates as the evidence accumulates.

The most demotivating property of self-directed preparation is invisibility. You study for six weeks and have no reliable way to know whether it worked. Progress Tracking exists to make improvement legible — and to make stagnation legible too, early enough to change what you are doing.

The rank ladder

XP earned through practice moves a student along seven ranks. The ladder is a motivational structure layered on top of measurement — it never substitutes for the mastery scores, which are what actually describe readiness.

◆ System 06

Learning Memory

Everything you have asked, answered and got wrong — saved, searchable, and carried forward. This is the system general AI tools do not have at all.

A conversation with a stateless chatbot is complete when it ends. Nothing about it informs the next one. Exam preparation is the opposite kind of activity: its entire value is cumulative, and a tutor whose memory resets between sessions cannot do the one thing a tutor is for — notice a pattern across time.

• Full session history. Every chat is saved and can be reopened. The explanation that finally made logarithms click is still there in March.
• Searchable. Find the question you asked three weeks ago without scrolling through everything since.
• Mistake records persist. Wrong answers from chats, drills and mock exams are retained as structured records, which is what makes longitudinal diagnosis possible at all.
• Context carries between sessions. The platform knows which exam you are preparing for, what you have already covered, and what you have historically struggled with — so you do not re-establish it every time.
• Memory is what makes the loop compound. Without persistence there is no ranking, no trend, no predicted score and no plan. Every other system on this page depends on this one.

◆ System 07

Personalized Learning

The adaptive layer that connects the other systems: your plan is rebuilt from your live diagnosis, not selected from a menu of preset tracks.

"Personalized" is used loosely across educational technology, usually meaning a student picked a level at signup. Here it means something narrower and more demanding: the platform's recommendation of what you should do next is derived from evidence it collected about you, and it changes when that evidence changes.

What is actually personalized

• What you practise — drill sets generated from your ranked weaknesses, not from a chapter order.
• What order you practise it in — highest score impact first, adjusted for how much time remains before your exam date.
• How a concept is explained — Zero adapts depth, framing and approach, and offers a different route when the first one does not land.
• Which language you receive it in — English, Arabic or Franco.
• Which exam everything is oriented around — the same topic is taught differently for the ACT's time pressure, the SAT's format and the EST's.
• What the platform stops showing you — mastered skills leave the queue, which is as important as what it adds.

Underneath the personalization sits a fixed, specialist-authored curriculum. The pedagogy is not improvised per student — the selection and sequencing of it is. That is the distinction between adaptive teaching and an AI making things up on request.

◆ System 08

Human Support

Real educators behind the content, real people behind the account — and a continuous improvement process that is human, not automated.

Human support is listed last, but it is the pillar the other seven rest on. The educational knowledge inside Si Math AI was built by experienced educators and exam specialists. They are not a support channel appended to an AI product; they are the source of what the AI delivers.

This is the practical meaning of the positioning statement: human experience is why it works. The AI is how the expertise is delivered, and it delivers it to far more students than any timetable would allow — but removing the educators from the loop would not leave a slightly worse platform. It would leave a chatbot.

◆ Summary

The eight systems at a glance