Overview

The AI feature that can analyse and predict student's future performance(e.g., results, behavior, and Final Grade Range) with dropout risk according to given data fields. This feature aims to help university to discover students that have a high risk of dropout so faculties can offer help and support in time. 

Use Case

In many current tertiary education institutions, student's and performance are always the main focus of every faculty. A student with lower CGPA score or performance will always catch the attention of the faculty member and allow the faculty members to focus on them. However, in actual cases, when a faculty discovers the student's performance is low. It is already too late when they take action as some students will have their own dropout probability and dropped out by the time.

In some cases, not just one but many student's performance and results are volatile. Causing the faculty members hard to expect what comes next, like even when a student is performing well, yet other alternate reasons can cause the student to dropout.

The Problem

A faculty was paying attention to the student that is not performing well lately. The student's results and performance have been dropping for the past 2 terms, meanwhile another student also seems to behave a bit down lately and results are going ups and downs. This gives the faculty member a hard time as it is hard to predict what comes next and in the end nothing is saved, both students dropped out.

The Solution

In this Case, AI Prediction can help to improve the efficiency of locating students with lower performance by projecting a student's performance in the future and benchmark it according to a specific condition stated. This allows the faculties to monitor the student's earlier and offer help or tutoring to the student based on the prediction.

The AI can also help to predict the student's probability of dropping out of the course. 

Problem Solved:

Predict Final Grade Range

Predict CGPA Range

Predict Student Behavior

Predicting Dropout Risk

Confidence-Weighted Decisions

Benchmark Comparison 

In short, AI is able to predict the student's future performance and benchmark to a specific standard based on configured condition.

Scenario

In this scenario, A student named John Doe have studied and completed 2 terms and was about to start term 3 soon.

Now, a faculty member wants to predict John's future performance and see if he have a high or low chance to dropout. Therefore, a few data has to be sent to AI for analysis and prediction use.

From Contact

From Individual Program Enrollment

From Individual Pathways

These are the data and fields to be sent to AI for analysis and prediction.

The button for the AI operation will be allocated at the Contact record page.

Clicking on the button will result in a payload display that helps user to double confirm the data to be sent to AI. Once it is confirmed, Click “Confirm”.

The AI will then begin analysing and predicting.

After a short while, a prediction result will be generated telling the predicted student's performance, result, and dropout risk. Then it will also benchmark to the condition set by the user, a confidence level is also displayed showing how confident AI is in this prediction.

The Results

Greatly improving the student's performance in time with right on time support. Faculty member's time is also saved 

Here is a sample of the Prediction:

Output and Benefits

The AI output provides a clear and professional summary that includes

• Predicted Final Grade
• Predicted CGPA Range
• Predicted Behavior
• Predicted Dropout Risk
• Confidence Level
• Comparison Benchmark

This feature provides prediction value across multiple areas, along with a confidence level and benchmark. This combination provides users not only the predicted result, but also how trustworthy the result can be. With a summary and benchmark also shows that how the performance of the student is and if the student is beyond or under average.

Estimated Cost - Per Request

For the latest pricing, please refer to GPT-4.1 API pricing (official)

Token calculation is based on OpenAI Tokenizer.

Request (Input)

Price (as of Jan 2026): $2.00 per 1,000,000 tokens
Calculated based on 2454 characters. 927 tokens

Input Cost = 927 x $2.00 / 1,000,000 = $0.001854

Response (Output)

Price (as of Jan 2026): $8.00 per 1,000,000 tokens
Calculated based on 3406 characters, 706 tokens

Output Cost = 706 x $8.00 / 1,000,000 = $0.005648

Input (Request) cost for 500 tokens (GPT-4.1)
500 tokens × ($2 / 1,000,000) = $0.001424

Total Cost: <$0.01