Services
Custom software developmentUmbraco developmentWeb developmentCRM developmentERP developmentAI & LLM integration.NET development & modernizationE-commerce developmentIntegrations & automationCloud & DevOpsMaintenance, support & project rescueDedicated teams & white-label
Our Work
Case studiesProducts
About Blog Contact
Web platforms · EdTech

Teacher-approved AI hints and a cleaner course model for TutorPro

AI hints that teachers approve before students see them, one sortable course list across databases and a sign-up flow that fits solo learners.

ClientTutorPro
IndustryEdTech
Year2025
CategoryWeb platforms
TutorPro landing page for language schools and tutors
The client

About TutorPro

Multi-tenant online platform for language schools, tutors and independent students.

TutorPro is a multi-tenant tutoring platform for language schools, tutors and independent students. Each school has its own tenant database, while shared content sits in a main database.

OSKI took on a set of improvements: making courses easier to find and manage, helping students who get stuck and fixing how independent students sign up.

The challenge

What had to be solved

Several of the problems traced back to how data was split between databases.

  • Courses sat in both the main and tenant databases, so sorting and filtering across them did not work well.
  • The course area had five tabs that were hard to use, and dashboards had no library widgets.
  • A student stuck on a hard exercise could only ask a teacher or a classmate.
  • Students who signed up on their own were placed in a school's tenant database.
The approach

How the team built it

One index for every course

OSKI added a CombinedCourses index to the main database. It pulls main and tenant courses together, so lists sort and filter correctly. Two tabs became a single Studio tab with a switch between public and school courses.

Dashboards that show courses

Students now see courses on their dashboard, and the library page lists each user's own courses.

Hints that guide, not solve

Azure OpenAI writes hints from the exercise type, the task and the correct answers. The prompt keeps the answer out of the hint. The student's own answers decide when a hint is needed.

Teachers keep the final say

Teachers can edit, regenerate, delete or approve each hint before any student sees it.

Fixes to everyday flows

Rescheduling a lesson now needs a reason and emails both sides. Homework appears as a table with lesson, student or teacher, deadline and status. Solo students register in the main database under a dedicated placeholder company.

The outcome

What changed

Students get help when they need it, and the data model is simpler to work with.

  • Courses from both databases appear in one list that sorts and filters properly.
  • AI hints reach students only after a teacher approves them.
  • Independent students are registered in a correct, resource-efficient way.
  • Every lesson change carries a stated reason and an email to both parties.
Screens

Inside the product

Start a project

Have a similar project? Tell us about it.

A spec, a Figma file or three paragraphs is enough to start.

  1. First reply · from an engineer, not a sales rep
  2. 30-min call · with Roman, our CTO, to clarify scope
  3. After the call · a written proposal with plan, team and price
Roman Oleksuk
Roman Oleksuk, CTOPrefer to talk? Pick a slot in Roman's calendar
We reply from a human mailbox. No newsletters. Your brief stays under NDA on request. We use your details only to reply; see our privacy policy.