pratium
TREN
Curriculum-linked • Human-reviewed • Learning evidence

Beyond memorisation.
Towards real learning.

More than a correct-answer count: see what you know, where you need support, and what you can do with a different problem.

Get started Explore the learning loop

AI supports. Teachers make the judgement.

Science

Grade 6 · FB.6.1.1

Beyond a score,
evidence of learning.

  1. Baseline assessment
  2. Guided practice
  3. Delayed post-test
  4. Transfer with different questions
  5. Teacher review

A correct answer with a hint does not replace evidence of unassisted learning.

Illustrative learning flow · Not live student data
Curriculum-linked questions Separate support records Delayed transfer Teacher review

The real learning loop

Finishing practice
is not the same as learning.

Assessment, practice and human review stay connected for the same learner and learning objective. The next step is not reduced to a test score.

  1. 01

    Measure the starting point

    A baseline assessment makes the learner's starting point visible for the selected objective.

  2. 02

    Support the thinking

    Hints, retries and the learner's explanation are recorded separately during guided practice.

  3. 03

    Assess again after a delay

    A delayed post-test is evaluated separately from supported success during practice.

  4. 04

    Try a different problem

    Transfer with different questions tests whether learning can be used in a new context.

  5. 05

    Review with a teacher

    Baseline, post-test and transfer results are considered through human review.

  6. 06

    Choose the next step

    The learning-objective state and reviewed prerequisites help guide the next learning step.

Connected learning tools

From generating questions
to following learning.

01

Curriculum-linked question bank

Content infrastructure based on Turkish Ministry of National Education sources, learning outcomes and objective mappings. Mapping status and questions requiring review are distinguished.

02

Varied questions and content

Multiple-choice and open-ended questions; charts, tables and visual content. Source-file tools are available according to the selected plan.

03

Guided practice with Prof. Prati

Choose an answer, retry with a thinking hint and explain your reasoning. Supported practice is kept separate from unassisted assessment.

04

Teacher tools

Assignments, live quizzes, class and learner reports, and reviews of objective alignment and learning gain within the teacher experience.

05

Learning Graph

Drafting and human-review workflows for prerequisite relationships between learning objectives. An AI suggestion is not an approved relationship.

06

Education Eval

Versioned benchmarks, blind human scoring and model comparisons. Historical assessments remain separate from new-version results.

One learner, a shared perspective

Personal for learners.
Clear for teachers.

Learners

Choose a topic, practise at an appropriate level, explain your reasoning and follow your progress.

Learner experience

Teachers

Review assignments, class progress and assessment results for individual learners.

Teacher tools

Parents

Follow progress through understandable summaries and support learning without reducing it to marks.

Parent experience

Institutions

Bring teacher and learner workflows into a shared framework. Define the scope of use and a pilot together.

Discuss a pilot

Evidence, with its limits visible

Real records.
A clearly defined scope.

A live-record check on 10 October 2026 verified completed pilot chains: baseline → guided practice → delayed post-test → transfer → teacher review.

3completed pilot cycles
1learner
2distinct learning objectives

This demonstrates that the learning workflow runs in production; it is not evidence of general attainment gains or a causal effect. The pilots cover Science and Mathematics. Science cycles moved from 40% at baseline to 100% on transfer; Mathematics scored 80% at both baseline and transfer.

Pilot timing detail

Timing note: In the first Science pilot, practice to post-test was 23.2 hours; the system's 24-hour gate ran from baseline. Both other pilots also exceeded 24 hours after practice.

Responsible AI approach

AI suggests.
Human review gives it meaning.

Quality checks

Independent model audits aim to reduce errors. Disagreements, visuals and uncertain mappings may require review.

Transparent measurement

Hint use, assessment results, objective links and human reviews are considered as separate records.

Data and permissions

Data-processing scope, access permissions and consent matter. Consult the application's current official privacy and terms pages.

Your questions, answered

How is this site different from pratium.com?

pratium.com.tr introduces the platform. Accounts, tests, assignments and learning tools are hosted on pratium.com.

Does English support include the application?

This introduction site is available in Turkish and English. Selecting a language here does not change the pratium.com application language; its existing language options apply.

Does a high score prove learning?

No. Support use, delayed assessment, transfer with different questions and teacher review are considered together. Completing a pilot does not establish general effectiveness.

Where can I find plans and prices?

Current plans, scope and prices are listed on the Plans page at pratium.com.

View plans →
How does an institution start a pilot?

Contact us to define goals, grades and subjects, permission processes and the assessment plan together.

Let's plan together

Choose the right start
for your institution.

Use a demo or pilot discussion to define subjects, teacher workflows and measures of success together.

Request a meeting by email ↗

This link opens your email application; it does not automatically submit a form.