AI in e-Devlet, MHRS and e-Okul: What Has Actually Changed?

Digital public services matter when they help people find the correct process and complete it with less uncertainty. However, an online appointment, document barcode or notification is not automatically an artificial intelligence application. This review, dated 18 October 2026, examines verified features of Türkiye's e-Devlet government gateway, MHRS appointment service and e-Okul education platform. It separates documented AI applications from possible future automation, helping businesses assess what software suppliers can reasonably promise.

Official sources describe AI within the Health Ministry's NeyimVar? application and set out an education AI policy. They do not establish that every transaction across these three platforms runs through one AI system. Nor do the reviewed sources provide a common, measured reduction in completion times. Claims about faster services therefore require a specific feature, date, measurement and comparison.

Verified Digital Services and Documented AI

e-Devlet: Document verification and account access

The official document verification service checks barcoded documents produced through e-Devlet. Its mobile application also offers QR scanning. For a business receiving a document, verification provides a more reliable starting point than judging its visual appearance. Whether the document satisfies a particular application remains a separate question for the receiving organisation.

Two-step login adds an account security step. A workflow should not depend on a customer giving an accountant, vendor or chatbot their government password. The user should control their official session. The business should request only the documentation and authority needed for the particular transaction, without turning convenience into shared account access.

MHRS: Appointments and confirmation rules

The MHRS service on e-Devlet provides an authenticated route to the official system. A Health Ministry announcement dated 13 May 2024 describes appointment confirmation and the intended reuse of released capacity. Such a process can follow explicit business rules; its existence alone does not prove the use of a generative language model.

Creating an appointment, confirming attendance and waiting to see a clinician are different measures. Users should check current requirements in their appointment screen or through ALO 182. A reminder can support attendance, but it cannot guarantee additional capacity or priority access. Suppliers should not present notification delivery as evidence that patients obtained earlier consultations.

NeyimVar?: A specific, documented application

The Health Information Systems Directorate's 28 December 2022 announcement describes AI in NeyimVar? and referral towards a suggested outpatient clinic through MHRS. The MHRS homepage still provides information about the application. This supports a limited claim about decision support and navigation. It does not establish that an AI model controls all appointments, and its suggestion does not replace assessment by a clinician.

e-Okul: Parent services and education policy

The MEB Student Information System is an official access route for authorised parents. The e-Okul page also displays school appointment and 2026–2027 registration links. These digital services do not, by themselves, show that generative AI determines school placement or student assessments.

The Ministry of National Education's June 2025 AI policy and action plan covers 2025–2029. A policy objective, a pilot and a feature available to every user are different stages. Reporting a new capability requires checking its name, availability, implementation date and conditions instead of treating a strategic document as proof of deployment.

Piyasadan.net / TurkeyOfficial Sector Analysis

Our commercial interpretation is that clear guidance around public services can be valuable without automating the official transaction. A small employer or supplier may first need an accurate document checklist, a named reviewer and a defined route for resolving omissions. Buying a complex agent before clarifying these responsibilities can leave the same confusion inside a more expensive system.

Consider a proposed supplier onboarding process. The business defines why a document is necessary. The supplier obtains it through the appropriate official service and submits it through an authorised channel. A reviewer checks it, then records the decision. An internal tracker can focus on reference number, responsible person, verification date and status instead of accumulating unnecessary personal details. This is a design example, not an existing Piyasadan government integration.

For teams outside Türkiye, accurate English explanations of Turkish service names can reduce misunderstanding. A TurkeyOfficial guide can explain where to find an official process. It cannot grant account permissions, access medical records or authorise collection of a child's information. Commercial value should come from understandable guidance and responsible task management.

Comparing Popular AI and Automation Systems

The following groups are proposed pilot candidates, not evidence that e-Devlet, MHRS or e-Okul uses these products. Before choosing a vendor, define the allowed data, contractual terms, review responsibility and response to errors.

Google Gemini & OpenAI ChatGPT Integration

A limited Gemini or ChatGPT pilot could evaluate how accurately each turns selected public instructions into plain-language guidance. Inputs should be current official material chosen by an editor. Outputs should remain drafts with source links. Passwords, identity documents, medical histories and student records are outside this example. Every additional requirement generated by the model should be checked against the official source before publication.

Claude & DeepSeek for Code and Text Automation

Claude and DeepSeek could be compared when preparing Turkish and English versions of an internal information guide. The assessment should test preservation of dates, exceptions and authority limits, not simply shorter writing. Development examples should use synthetic records. A useful requirement is that missing evidence triggers referral to a reviewer rather than an invented answer or an attempt to enter a government account.

n8n, Make and Zapier for B2B Workflows

A proposed n8n, Make or Zapier workflow could monitor document review dates inside a business's own authorised system. Missing information would create a task for the responsible person; completion would close it. Any connection to a public platform requires a separately verified, permitted access method. A publicly visible website does not establish an open API or permission for bulk processing. No appointment-catching service or automated account access is proposed here.

Building a Safe Pilot With Measurable Results

KVKK's AI recommendations emphasise limited purposes, proportionality, security and protection by design. Its generative AI guide is another reference for examining privacy implications. A real deployment needs qualified review of its legal basis, retention periods and access controls.

An initial test could use twenty questions containing no personal information. Record correct destination links, outdated instructions and omitted conditions. Measure the time needed for human review as well as answer generation. To claim improvement, compare equivalent tasks before and after the pilot. Faster output with more incorrect directions should fail the evaluation.

Include difficult examples: an expired announcement, a missing source page and a question outside the guide's scope. The expected response should identify the limitation and direct the user to official assistance. A clear handover is a useful outcome; a confident but unsupported instruction is not. Keep a dated record of approved answers so future editors can identify what changed.

Frequently Asked Questions

Are e-Devlet, MHRS and e-Okul entirely operated by AI?

The reviewed official material does not support that general conclusion. Digital transactions, rules-based automation and specific AI-supported applications should be assessed separately.

Can an AI assistant guarantee an earlier MHRS appointment?

This review verifies no such authority or guarantee. Check appointment availability through official MHRS channels and distinguish reminders from guaranteed access.

Should a parent upload e-Okul records to a public chatbot?

The proposed use does not require exporting student records. General guidance should work without personal details and direct the parent to the authorised service.

What should a business request from an automation supplier?

Ask for the data scope, permitted connections, human approval steps, retention rules and evaluation method. A working demonstration does not prove authority to act for a public institution.

Conclusion and Commercial Opportunities

Useful public-service automation begins with accurate guidance and clear responsibility. Explanation, translation and internal task tracking provide bounded areas for evaluation. Readers of TurkeyOfficial and Piyasadan.net can compare suppliers by checking authorised access, data limits and measured results before considering product names or expansive claims about AI.

Sources and Review Note

This article was prepared with AI assistance using primary sources.

Sources checked on 18 October 2026. Historical announcements describe the dated arrangements; current transaction conditions should be checked through the relevant official service. Workflow examples are editorial proposals.

1. e-Devlet — Document verification.
2. e-Devlet — Two-step login.
3. e-Devlet — MHRS service.
4. Ministry of Health/Burhaniye State Hospital — Appointment confirmation, 13 May 2024.
5. Health Information Systems Directorate — NeyimVar?, 28 December 2022.
6. MHRS — Official homepage.
7. e-Devlet — MEB Student Information System.
8. MEB — Official e-Okul page.
9. MEB — Education AI policy and action plan, 2025–2029.
10. KVKK — Recommendations on personal data protection in AI.
11. KVKK — Generative AI and personal data protection guide.