Save timeImprove quality
AI HR assistant
Handle employee queries, automate routine HR processes and free your HR team to focus on the work that actually requires human judgement.
AI for education: use cases, governance and how to start
Learning systems now have memory, foresight, and feedback loops. Personalized education at scale is no longer a vision — it's a data-driven reality.
In short
In education, AI is used for study assistants that answer at any hour, personalised learning paths, feedback on exercises, content creation for teachers and support for administrative and student services. It helps when pedagogy stays in charge: the institution decides the objectives and the content, and the AI adapts pace and format. Uses that decide admission, grade students or monitor exams are high-risk under the EU AI Act and need the strictest controls.
Learning that adapts.
- Personalized learning paths and content delivery
- AI-based tutoring and adaptive assessments
- Student performance prediction and intervention
- Automated grading and feedback generation
- Virtual assistants for student support
- Curriculum planning with analytics insights
Solution plays
View all solutions
Save timeImprove quality
AI learning & development
AI-powered learning that personalises content, tracks skill gaps and keeps your workforce ahead of what the business needs — without the overhead of traditio…
Save timeImprove quality
AI onboarding assistant
Guide every new customer, employee or partner through onboarding with an AI assistant that adapts to their context, answers their questions and keeps them mo…
Save timeImprove quality
AI workforce
Deploy AI workers that handle defined roles end to end — from customer-facing interactions to back-office tasks — with the consistency, availability and spee…
The European angle: EU AI Act and sector rules
The EU AI Act (Regulation (EU) 2024/1689) lists education in Annex III. AI systems used to decide access or admission, to evaluate learning outcomes, including when that evaluation steers the learning process, to assess the level of education a person will access, and to monitor or detect prohibited behaviour during tests are high-risk. They need risk management, data governance, human oversight, logging and transparency towards those using them. Annex III obligations apply from December 2027, after the Digital Omnibus on AI entered into force on 27 July 2026; AI literacy (Article 4) and Article 50 transparency already apply. Check the consolidated text on EUR-Lex or the AI Act Service Desk.
One practice is banned in education and that ban already applies: emotion recognition systems in educational institutions, except for medical or safety reasons. AI literacy duties also apply since February 2025, which means teachers and staff who use AI should understand what it does and where it fails. A study assistant that explains content or answers questions about a course is usually not high-risk, but it becomes so if it starts grading students or deciding their path.
Student data deserves particular care. Many learners are minors, and the GDPR requires special protection for children’s data and clear information in language they can understand. Educational platforms and content must also meet accessibility requirements, and public institutions buy through public procurement rules. This is practical guidance, not legal advice.
How to start
- Separate use cases by risk: support and content creation on one side; admission, grading, level placement and exam proctoring on the other, with their own governance.
- Start with a study or support assistant that answers only from the institution’s own materials, cites them and refers to a teacher or service when it cannot help.
- Agree with teaching staff on what “good” looks like: learning objectives, the tone of feedback and the limits of what the assistant should do for the student.
- Pilot with a course or cohort, review answers with teachers and measure use, satisfaction and impact on the workload of support teams.
- Before any use that evaluates students, document the classification, test for bias across groups and design how a teacher reviews and a student can challenge a result.
Success stories in this sector
Thinkia products that fit
Decisions you will face
Frequently asked questions
Is AI-assisted grading allowed?
Yes, but AI that evaluates learning outcomes is high-risk under the EU AI Act. That means human oversight, documented testing for accuracy and bias, logging and clear information for teachers and students. In practice, the AI proposes and the teacher decides.
Can we use AI to proctor online exams?
Monitoring and detecting prohibited behaviour during tests is a high-risk use under Annex III, and inferring students’ emotions in educational settings is prohibited. Any proctoring system needs a careful legal and data protection review before deployment, with a clear alternative for students.
Does a virtual study assistant count as high-risk?
Usually not, if it explains content, answers questions about the course and guides the student. It must make clear that it is AI. It changes category if it starts deciding a student’s level, path or grade.
How do we keep pedagogy in charge?
By grounding the assistant in the institution’s own materials, setting the objectives with teachers, defining what the assistant must not do for the student and reviewing samples of its answers regularly. Thinkia built an EdTech platform with a real-time virtual study assistant and personalised learning paths on that principle.
What should we check with student data?
Legal basis, minimisation, where data is processed and whether providers can use it to train models. With minors, information must be clear and age-appropriate. Security and access controls for learning analytics matter as much as the model itself.
Further reading
Key terms
Want a solution mapped to your context?
Talk to an AI Expert