ChatGPT for Business-Oriented Predictive Models: My Experience as a Data Scientist

Leveraging AI Prompts and Interactive Machine Learning

Introduction

The ideal scenario is as follows: if I am a decision-maker in a company, assigning a task to a robot should feel like delegating it to a skilled data scientist. Despite lacking background knowledge of the project, the robot should be able to learn, ask questions, make suggestions, and execute the task. This includes writing code, running simulations, generating reports, and managing the model’s operation and maintenance. This is the true potential of GPTs.

As discussed in this post, the pinnacle of AI utilization is an interactive system that can take on assigned tasks, such as building predictive models to achieve business goals. Current GPTs can handle many tasks and significantly reduce the need for human labor. For instance, I can upload an image of a spreadsheet and have GPTs assist with digital conversion, perform basic summary calculations, and even analyze reports. I have utilized these capabilities in my data analysis work. However, as a data scientist building predictive models, especially business predictive models, I haven’t found an ideal modeling GPT. In addition, I suspect that even if such GPTs exist in the future, they will find…