Impressit logo

Welcome to impressit

burger iconMenu
Back to Case studies
global-study product logo

AI assistants for students, colleges, and education agencies

We made

AI assistants

Industry

Education

Technologies

GPT 4.1, FastAPI, React, React Native, GitHub, Lambda, DynamoDB

Founded

2020

Location

United Kingdom

Partnership period

January 2025 — Present

Team size

Cosmo — 1

Julia — 4
Global Study Hub — 4

Executive Summary

Impressit was involved in the fine-tuning and development of three  AI assistants for our client, GlobalStudy:


  • Cosmo. Our team improved the chatbot's response relevance and speed, transforming the tool into a reliable asset for agencies. 
  • Julia AI. We created this AI assistant from scratch. Julia's goal is to strengthen universities' ability to engage and convert prospective students. 
  • Global Study Hub. This is a groundbreaking solution under development to empower students with direct access to university recommendations.
Product image
Product image

Client Background

GlobalStudy, an organization dedicated to supporting students in gaining admission to universities abroad, sought to expand and enhance its digital offerings. The company creates AI-driven products for international student recruitment agencies, colleges/universities, international students, and service providers.

Challenges and Objectives

GlobalStudy faced challenges regarding the quality and speed of its product development. The client required a solution that improved not only product performance but also delivery timelines. 

Impressit was engaged to address these issues through the development and fine-tuning of three distinct AI-driven products.
 

Cosmo 



Cosmo is an AI chatbot designed for agencies that assist students in applying to international universities. It provides personalized responses to student queries based on information uploaded by the agencies, including admission costs, deadlines, and other critical data. 
For example, a student might ask, "I have a budget of X and want to study computer science in Germany. What are my options?" Cosmo processes this query and delivers appropriate recommendations based on pre-loaded datasets such as PDFs, Excel files, and other documents. 

Challenges addressed: 
Initially, Cosmo delivered inaccurate and low-quality responses, limiting its usefulness. Its task to generate insights from seed memory data was insufficiently executed, leading to unsatisfactory outcomes for both students and agencies. 

 

 

 

 

Julia 



Unlike Cosmo, Julia AI is designed specifically for universities to engage prospective students directly. It is integrated as a widget on university websites, functioning as a virtual assistant that answers questions while simultaneously gathering valuable student data. 
How does it work for universities? 

  • Captures key student information (e.g., email addresses, phone numbers, level of interest). 
  • Utilizes this data to populate a lead dashboard, which allows university sales teams to contact students, answer follow-up questions, and promote their offerings effectively. 
  • Enables customization, such as tailoring greeting messages and prompts to reflect the university's tone.

Challenges addressed:

The development team encountered technical challenges due to the complex integration of CSW signs, PDF documents, and the GPT model used at the time, which had limitations in processing various file types. To overcome these issues, a former Google engineer with expertise in startups was brought onto the team. 

Global Study Hub  


The Global Study Hub is a web-based solution currently under development. Its purpose is to streamline the admissions process by eliminating the need for intermediaries such as agencies or university representatives. 
Students will be able to interact directly with the platform, asking questions, taking notes, and receiving recommendations for educational institutions. This direct approach will simplify the user experience while ensuring data accuracy and transparency.

Product image

Value Delivered

By collaborating with GlobalStudy, Impressit delivered measurable results that addressed critical pain points across all three products. 


The relevance of Cosmo's responses dramatically improved, increasing accuracy from 18.75% to 92.86%. We enhanced answer quality and response speed through meticulous model fine-tuning. 



Leveraging our expertise in generative AI, we created Julia from the ground up. We integrated features to enhance lead generation capabilities and improve data gathering functionality. Impressit also added customization options, enabling universities to adapt Julia AI to their specific requirements while maintaining compatibility with their existing systems.

 

As a result, Julia's key possibilities include:


 

  • Students can ask up to five questions for free before transitioning to the subscription model ($2/month)
  • Notes and notifications can be added to responses, serving as reminders for essential application tasks.
  • Data is sourced from a broad spectrum, ranging from web links to inaccessible documents, such as CSVs and PDFs, to ensure comprehensive recommendations.



Finally, four of our developers are actively working on the Global Study Hub, focusing on end-to-end development to create a seamless and intuitive user experience.

 

Future Outlook

Global Study continues its mission to revolutionize international student admissions with cutting-edge AI technology. Impressit remains a dedicated partner, ensuring their products align with the evolving needs of students, agencies, and universities.

 

 

 

Say hello to our team

Write to us in the chat or complete contact form, and we will contact you within 24 hours.