Skip to content

Clinical learning intelligence

Know exactly what you don’t know.Then study only that.

Si MedIQ reads your own lectures, watches how you answer, and names the specific confusion behind each mistake. Every day it hands you one thing to do, how long it takes, and the reason it won — computed from your evidence, not a generic question bank.

Deterministic mastery model · Built from your material and your exam dates · No generic question bank

  1. Question · cardiology

    Three weeks post-MI, recurrent monomorphic VT. Which agent is contraindicated?

    ABCD
  2. Si is reading the answer

    • Chose Dsotalol — a class III agent
    • 41 s2.3× your median on this topic
    • Marked “certain”confident and wrong
    • 3rd timesame two classes, same swap
  3. Confusion detected · 3 pieces of evidence

    Class Ic vs class III

    The post-infarct risk you are describing belongs to the class Ic agent, not the class III one. That is not a knowledge gap — it is two classes crossed.

  4. Mastery recomputed

    Vaughan Williams classification0.33 → 0.29
  5. Your next best action

    Fix one confusion: class Ic vs class III

    10 min · 3 questions · chosen over 11 alternativesStart

A worked example. In the product every number here is computed from your own attempts.

The problem with doing more questions

You did four hundred questions. You still don’t know which sixty will fail you.

A question bank can tell you how many you got right. It cannot tell you why you got one wrong, whether you would get it wrong again, or which of the two hundred things you have studied is the one about to cost you the exam.

Your average across the bank

74%

Every attempt counted once, forever, at equal weight. It went up because you answered more, not because you understand more.

Your mastery, as Si computes it

0.41

Weighted by question difficulty, decayed by time, penalised when you were confident and wrong, and held down by an open confusion.

One is a scoreboard.
The other is a prediction.

Inside one answer

Four things are true about every answer. Most products record one of them.

Reading one submission
Which option
D — sotalol
Not just right or wrong. Each distractor is written to catch one specific way of thinking, so the one you picked names the error.
How long
41 s · 2.3× median
Measured against your own median on that topic, not a global average. Slow and right is a different state from fast and right.
How sure
Marked “certain”
Asked after you commit, never before. Confident and wrong is the most useful signal in the system, and it is the one nobody else collects.
What came before
3rd time
One mistake is noise. The same mistake in the same direction three times is a confusion — and nothing is named until the evidence supports it.

Exam intelligence

The same weaknesses, in a different order, because the date moved.

Urgency is not a label you apply to a topic. It is what falls out when weakness, importance, how overdue a review is and how close the exam is are scored together. Move the date and the ranking moves with it.

30 days out

  1. 1Renal tubular transportnever practised
  2. 2Vaughan Williams0.29 · confusion open
  3. 3Acid–base compensation0.44
  4. 4Cranial nerve palsies0.71 · due

Coverage wins. Your time goes to what has never been measured at all.

6 days out

  1. 1Vaughan Williams0.29 · confusion open
  2. 2Acid–base compensation0.44
  3. 3Cranial nerve palsies0.71 · due
  4. 4Renal tubular transportnever practised

Risk wins. An open confusion on an examinable topic outranks a subject you have not opened.

The loop

Nothing here is a one-off. Every answer changes what you are shown next.

  1. 01

    You answer

    One question, and how sure you were. That certainty is the part every other product throws away.

    C · 41 s · “certain”

  2. 02

    Si reads it

    Which distractor, how long, how confident — and whether this is the third time the same idea has gone wrong.

    4 signals read

  3. 03

    The model moves

    Mastery is recomputed from evidence rather than tallied, and a confusion is opened or closed.

    0.33 → 0.29 · confusion opened

  4. 04

    The next sighting is set

    Right but unsure comes back sooner than right and certain. Failing something you had mastered comes back tomorrow.

    due in 2 days

and the next answer starts it again

Your material, not a catalogue

It only ever asks you about what you were actually taught.

Upload a lecture and Si MedIQ reads it, pulls out the concepts, works out how they relate, and writes questions against those concepts — citing the slide each one came from. Nothing in your mastery score comes from a bank someone else wrote for a different curriculum.

  1. 01ReadSlides, notes and PDFs become a structure: concepts, their hierarchy, and the source reference for each one.
  2. 02QuestionItems are written against your concepts, with distractors that each test one specific way of going wrong.
  3. 03MeasureEvery attempt updates mastery for the concept it belongs to, never for a subject as a whole.
  4. 04ExplainThe tutor answers with your material in hand and shows the line it is working from — or says it does not know.

Start with one lecture.

Upload one set of slides and add the date of your next exam. Si MedIQ reads the material, builds the concept structure, writes questions against it, and has a plan for your week before you have finished making coffee.

Every confusion you close is one you don’t carry onto a ward.

Free while you are studying · Your material stays yours · Delete everything in one click