Updated 10 October 2026.
Medical revision platforms increasingly use similar language for different things. A knowledge profile might estimate proficiency. A knowledge tutor might schedule repeated recall. An AI tutor might discuss the reasoning behind a question. All three can be useful, but buying one because its name resembles another is a poor way to choose a learning tool.
The distinction matters when you are struggling. You may need to find an overlooked topic, remember information more reliably or understand why two plausible answers are not equally appropriate. Those are different problems. A system that solves one elegantly may do little for another unless the learner chooses the next activity deliberately.
A profile describes or estimates; a tutor should help you change something
A profile is a representation of the learner based on selected evidence. It can organise observations, estimate an ability or identify where more information is needed. It does not become teaching merely because its output is personalised.
AMBOSS describes Knowledge Profile as a question-based model of topic and competency proficiency. Its role is different from that of a conversation designed to explore why the learner selected a particular answer.
A useful way to distinguish the functions is to imagine a recurring mistake. The profile identifies the area in which the mistake occurs. A tutor explores the learner's interpretation. A later assessment checks whether the explanation has changed what the learner can do. None of these steps should silently stand in for all the others.
The critical question for AMBOSS is therefore not whether its profile is granular. It is whether the associated next action addresses the mechanism of the error. A very precise label can still lead to a relatively generic intervention.
What does Passmedicine mean by Knowledge Tutor?
Passmedicine's homepage describes its Years 1-3 Knowledge Tutor as Anki-style. That description points towards structured recall and repetition, rather than establishing that the feature is a free-form conversational tutor or an exam-score forecasting model.
It would be equally mistaken to assume that the word tutor proves the presence of a large language model or that a recall-based tool is educationally inferior because it is not conversational. The relevant question is whether the user needs a structured way to revisit material or a dialogue about an unresolved distinction.
For example, someone who can explain a concept accurately but fails to retrieve an associated fact later may need a different activity from someone who recalls all the facts but draws the wrong conclusion. The first problem might suit focused recall practice; the second may require examining the reasoning process.
The exact Passmedicine resource matters. A feature described for early medical years should not be assumed to operate identically across every examination package. Current product descriptions should take precedence over a general impression formed from the brand name.
What iatroX's Socratic Tutor is designed to do
The iatroX Tutor page describes a question-linked conversation: the learner attempts a question and then discusses it within the relevant examination context. The Tutor can explore reasoning or give a direct explanation; learners can save highlights, create revision cards and download the conversation. It is not presented here as a substitute for every other form of revision.
The useful element is making the learner's reasoning visible. Suppose a candidate answers a question about the next step by naming the likely diagnosis. A simple explanation of the diagnosis may not address the error. The learner may already understand the condition but have answered a different task from the one posed.
In that hypothetical example, an effective discussion would ask what the question requested, which options answer that request and what information changes the priority. The aim is not to prolong the conversation. It is to isolate a distinction the learner can use later.
A concise exchange that ends with a clear correction may be more useful than a long dialogue that merely paraphrases the rationale. Socratic teaching should not become a ritual of withholding a straightforward answer when the learner genuinely needs one.
AMBOSS and Pastest also offer AI support
AMBOSS's AI Mode for Learning supports questions, study-material uploads and recommendations linked to its resources. Pastest's MRCP offering describes AI tutoring with hints and explanations. iatroX is therefore competing in a category with overlapping capabilities, not introducing conversation to an otherwise non-interactive market.
That makes implementation more important than the label. Does the tool retain the relevant question and examination context? Does it ask about the learner's actual reasoning? Can it make an explanation clear without implying that its confidence is evidence of correctness? Does the learner have a sensible next activity after the discussion?
Source grounding and appropriate uncertainty also matter. A tutor should distinguish a misunderstanding of an examination convention from a question requiring current clinical guidance. These are proposed evaluation criteria for any tutor, including iatroX, rather than claims that one company's product handles every case perfectly.
Uploading another provider's copyrighted question is not a necessary condition for this workflow. A learner can describe the underlying concept in their own words or practise an independently provided question on the same topic. No automatic AMBOSS-to-iatroX profile import is implied.
The quiet problem: help can conceal what you still cannot do
There is an obvious difference between choosing an answer unaided and choosing it after several hints. A platform may support both activities legitimately, but the learner should know which one occurred. Otherwise, assisted success can become a misleading proxy for independent performance.
The same risk exists with fluent explanations. Recognising that an explanation makes sense is not identical to generating the reasoning yourself. A good tutoring session should therefore end with something the learner can state or apply without simply repeating the tutor's wording.
One practical approach is to give a brief account of the deciding distinction, then consider a changed version of the problem. What information would make a different answer appropriate? Which tempting option was rejected, and why? These are suggested prompts for evaluating understanding, not a validated scoring instrument.
A later, fresh question offers a cleaner test than immediately repeating the identical item. The goal is not to eliminate help from learning. It is to avoid confusing the conditions under which help is useful with the conditions under which independent performance is being assessed.
Repetition, explanation and planning have separate roles
Anki's FSRS documentation illustrates the scheduling side: review history informs when cards should return. That is different from exploring a misconception in conversation, and different again from deciding how to distribute study time across an examination syllabus.
The iatroX Study Plan provides a place to organise activity, while its question banks and Tutor address practice and explanation. The productive comparison with Knowledge Profile is whether those activities lead to an appropriate next step, rather than whether every platform attaches the same name to its features.
A learner does not need to use every mode in every session. Sometimes a short recall activity is enough. Sometimes an explanation is the highest-value intervention. Sometimes the important action is a mixed block that tests whether the recently discussed principle can be recognised without a cue.
Where the next stage of personalisation should go
iatroX's direction is to connect gap identification, question practice, tutoring and later review more closely, while extending continuity into professional learning. That does not mean every conversation should automatically raise a mastery score. Assisted performance and unaided performance should remain distinguishable as these systems become more connected.
The broader opportunity is a learning history that remembers what was difficult and what the learner did about it, without pretending that activity alone proves competence. AMBOSS's learner model, Passmedicine's recall-oriented tools and conversational tutors all contribute ideas to that future.
For the individual learner, the decision remains practical. Choose the tool that addresses the problem in front of you, then look for evidence that the problem has actually changed. The most useful tutor is not necessarily the one that talks most, and the most useful profile is not necessarily the one with the most detailed colour coding.
Frequently asked questions
Are Knowledge Profile and Knowledge Tutor different names for the same feature?
No. In the examples discussed here, a profile models or summarises knowledge, while a tutor supports a learning activity. The exact product description matters more than the similarity of the names.
Is a conversational AI tutor always better than spaced repetition?
No. Conversation and scheduled recall address different needs. A learner who needs to retrieve selected material more reliably may not benefit from replacing every short review with a dialogue. A learner with a persistent misconception may need more than another repeated prompt.
How should I use iatroX Tutor after getting a question wrong?
State your reasoning, identify the point at which it diverged from the explanation and check whether you can express the corrected distinction yourself. Then use an appropriate fresh task to assess application rather than treating the conversation alone as evidence of mastery.
Explore the reasoning behind your next question with the iatroX Socratic Tutor.