What Is AI Coaching? How durumum.net Artificial Intelligence Helps You Prepare for YKS

A student in the second year of university prep — targeting an equal-weight program, working with a private tutor once a week — started every session the same way: "How was this week?" She was never sure what to say. "Fine" or "I had trouble with math." Her tutor gave general advice. The session ended. Neither of them knew which specific topic from that week's practice exam had caused the problems, or why the same mistakes kept appearing. Coaching ran on intuition, not data. AI education coaching is where that changes.

How AI Coaching Differs from Traditional Coaching

Traditional coaching is human and valuable — an experienced teacher beats AI on motivation, discipline, and contextual judgment. But it has a structural constraint: it operates on memory and intuition rather than data. Tracking which topic a student got wrong and when, across dozens of exams, is not practical in a weekly session.

AI coaching fills that gap. Four concrete differences:

1. Everything Is Data-Driven

Once exam results are entered, AI analysis works from numbers rather than impressions. Not "you have trouble with math" — rather, "across your last three practice exams, you averaged 2 out of 4 on Probability questions and 6 out of 8 on Geometry." That distinction completely changes how you allocate study time.

2. Available Around the Clock

At 1 AM after finishing a practice exam, when you want to review your priorities for the week, your tutor is not available. AI analysis is always there. You can enter exam data and check updates at any hour of any day.

This is especially important during intensive exam periods. Students entering multiple practice tests within three days benefit from real-time updates that allow rapid priority adjustments.

3. No Bias

Even the best human coaches can be influenced by a student's social presence — the perception that someone is working very hard can soften the assessment of where they are actually falling short. AI analysis does not have that. What the data shows is what gets reported.

4. Real-Time Updates

Each time new exam data is entered, the analysis updates immediately. With a human coach, last week's data gets summarized in the following session. AI has no such delay.

durumum.net AI Features

The AI coaching component in durumum.net produces three core outputs:

Weak Topic Detection

After exam data is entered, the platform calculates where performance is lowest across TYT and AYT subjects. But this goes beyond a general "math is weak" assessment. Where the data allows, it breaks down by topic: in Turkish, is the issue reading comprehension or grammar? In math, is it Trigonometry or Limits?

That specificity matters because study time is limited. "Study math more" is far less useful than "put four hours this week into Trigonometry and Limits." The second version uses time much more efficiently.

Study Plan Suggestions

AI analysis does not stop at identifying weak areas. Looking at how many days remain until the exam, where marginal gains are highest, and what past performance trends look like, it produces a priority ranking for the current week.

Example: a student with six weeks left can probably bring TYT Social Studies from 15 to 20 correct, but is well below average for quantitative-track students in AYT Math. The AI suggests spending less time on Social Studies and more on Math — because the Math improvement translates into far more points.

Score Projection

Based on current performance trajectory, the platform shows what score range might be achievable by exam day. This projection is not guaranteed — exam-day conditions, stress, and question distribution always vary. But seeing "at this rate, my TM score could land in the 380-400 range" gives a realistic basis for program selection decisions.

The score projection also functions as an early warning system. If correct counts in a particular subject have been declining over three consecutive exams — despite continued studying — the AI flags it. Maybe the resource is wrong, maybe the topic sequence is off; noticing requires looking at the data.

A Real Usage Scenario

To make this concrete: consider a student targeting an equal-weight program. Her TYT Turkish and Math scores are solid, but her AYT Literature score is inconsistent — sometimes 18, sometimes 11. She interprets this as a knowledge gap and starts working through the topics from the beginning.

After six months of exam data is entered into durumum.net, the AI analysis says something different: Literature scores are low on high-stress exam days and high on low-pressure school practice tests. This finding points not to the topic but to exam strategy. The issue is not knowledge — it is a habit of skipping questions under pressure.

Another example: a student who has been stuck in the 22-24 correct range on TYT Math for months. The tutor says "solve more problems"; the student does. The number does not move. Looking at the data on the platform reveals: the student gets easy and medium questions almost perfectly but skips two-thirds of the hard questions. The issue is not speed — it is hard-question strategy.

The Limits of AI Coaching

Expectation management matters. Here is what AI coaching cannot do:

• It cannot generate motivation. On tired, unmotivated days, pushing yourself to work requires human support. AI provides analysis; it does not encourage.
• It does not explain topics. It can identify that Trigonometry is a weak point, but it does not show how to study it. That still requires resources, video lessons, and practice.
• It does not work without data. Without at least a few exam results entered, meaningful analysis is not possible. Getting real value from the platform requires actual input.
• It cannot read context. "This week there was an illness in my family and I could not focus" is a factor AI does not see. External events can skew the data. Reading that context is your job.
• It cannot guarantee results. Score projections are estimates, not guarantees. Exam-day conditions — question distribution, stress level, health — are always variable.

Privacy and Security

AI coaching requires the platform to process your data — which naturally raises the question of where that data goes.

On durumum.net, your exam data and personal information are used only for analysis purposes. They are not shared with third-party advertising networks, data brokers, or educational institutions. Your national ID number is used only to retrieve exam results automatically and is not stored on our servers.

If you delete your account, your data is deleted with it. No profiling is done without explicit consent.

Who Benefits Most from AI Coaching?

• Students who take practice exams regularly: Analysis deepens as data accumulates. One practice exam per week means at least four data points per month. Trend analysis becomes meaningful at this frequency.
• Students struggling with field selection: TM or MF? Quantitative or verbal? Making this decision based on which field you are improving in fastest — and actual score-type weights — is far healthier than intuition.
• Students with tutors or private teachers: Sharing AI analysis in those sessions makes them concrete. "I had trouble this week" becomes "my Trigonometry score has dropped in the last two practice exams, here is the chart."
• Students prone to panic: The feeling before exams that "I haven't studied anything, I don't know anything" is very common — but usually does not match the data. Seeing your actual situation through the platform partially relieves this. Data-based self-assessment is a much healthier starting point than intuition-based panic.
• Independent learners: Students who do not attend coaching centers and plan their own studying are most critically dependent on AI coaching. Without a weekly priority list and subject-level feedback, independent studying can easily lose direction.

How to Get Started

Using AI coaching starts with setting up exam tracking:

1. Create an account — email and password, takes under five minutes
2. Verify with your national ID and student number — past exam results load automatically
3. See your first analysis — subject-level performance and score calculation on the analysis screen
4. Let data accumulate — update after each new practice exam, and AI suggestions sharpen over time

No payment or subscription is required to start. Core analysis and AI coaching are completely free.

In exam prep, knowing what you do not know is harder than knowing what to study. Create an account, enter your exam data, and look at the analysis screen — which areas are genuinely causing problems becomes much clearer when the numbers do the talking.