Ghidul tău USVT – AI Student Guide

• 

  Ghidul tău USVT – bilingual, source-grounded AI student guide

• 

  Source-grounded answer with a direct link to the official USVT page

• 

  Bilingual English experience with source-grounded USVT answers

Inspiration

Finding accurate university information should not require searching through dozens of webpages, PDFs, and administrative documents.

Prospective students often need quick answers about admissions, study programs, tuition fees, campus services, and student life, but official information is frequently distributed across multiple sources.

We created Ghidul tău USVT – AI Student Guide to demonstrate how AI can transform verified institutional information into a conversational, transparent, and easy-to-use experience while always keeping the official university source at the center.

Although the first implementation was developed for the "King Mihai I" University of Life Sciences from Timișoara (USVT), the architecture was designed as a model that could be adapted by other higher education institutions seeking to modernize student communication.

---

What it does

Ghidul tău USVT is a bilingual (Romanian–English) AI-powered student guide that helps prospective and current students quickly access official university information.

Instead of navigating multiple webpages, users can ask natural-language questions or explore essential topics such as:

• Admissions
• Faculties
• Study programmes
• Campus services and student life

Each response is concise, easy to understand, and linked to the official university source whenever available. The assistant is intentionally designed to avoid presenting uncertain or unsupported information as factual, encouraging users to verify administrative decisions using the official documentation.

---

How we built it

The application was developed as a lightweight responsive web application using HTML5, CSS3, and modern JavaScript.

Its knowledge base is organised inside a dedicated data module, allowing official information to be reviewed, updated, and maintained independently from the application logic.

The conversational experience, multilingual interface, accessibility improvements, testing, documentation, and iterative refinement were developed during OpenAI Build Week using OpenAI GPT-5.6 and Codex.

The project is open source, version-controlled on GitHub, and publicly deployed on Vercel.

---

Challenges

The biggest challenge was not building a chatbot—it was building a trustworthy educational assistant.

Official university information is spread across multiple webpages and changes over time. Our objective was to transform that information into concise answers without losing transparency, context, or confidence in the original source.

Another challenge was designing a bilingual experience that remained intuitive, accessible, and consistent across desktop and mobile devices.

---

Accomplishments

• Functional Romanian–English AI student guide
• Source-grounded responses connected to official university information
• Clear distinction between verified information and administrative recommendations
• Responsive interface optimised for desktop and mobile
• Modular architecture for future content expansion
• Public deployment on Vercel
• Open-source GitHub repository
• Demonstration of a practical AI solution for higher education

---

What we learned

Building an educational AI assistant requires much more than generating fluent answers.

Trust is created through transparent sources, carefully structured content, clear limitations, responsible AI behaviour, and an architecture that institutions can easily maintain over time.

We also learned that conversational AI can significantly simplify access to institutional information without replacing official university communication.

---

What's next

The current prototype represents the foundation of a broader educational AI platform.

Future development includes:

• expanding the official knowledge base;
• automated synchronisation with approved university sources;
• multilingual support for additional languages;
• improved accessibility and usability testing;
• analytics for frequently asked questions;
• secure retrieval mechanisms connected exclusively to verified institutional content;
• an architecture that can be adapted by other universities and educational institutions.

Our long-term vision is to help higher education institutions deliver faster, more transparent, and more accessible student support through trustworthy AI.

Built With

• artificial-intelligence
• bilingual
• codex
• css3
• education
• education-technology
• github
• higher-education
• html5
• javascript
• openai
• responsive-design
• student
• student-support
• vercel