Introduction

ChatGPT introduces an exciting new feature: the ability to extract essential information from uploaded text files like PDFs, TXTs, PPTs, DOCs, XLSXs, and CSVs. This analysis capability simplifies work by allowing users to focus on critical tasks. Learn more about supported file formats and limitations at Advanced Data Analysis.

Prerequisites

An active Plus or Enterprise subscription is required to use this feature.

Configuration Steps

The feature is in phased deployment, so you may need to follow a few steps to enable it. Here’s how:

1. Log in and switch to the GPT-4 model. If you see the file upload icon (a small plus), proceed to the next step. If not, access Settings & Beta to activate the feature.
2. In the sidebar, click the context menu by your account name and select Settings & Beta.
3. In the Settings modal, find Advanced data analysis under Beta features and enable it.
4. Return to the GPT-4 interface. If the file upload button isn't visible, hover over the GPT-4 label, select Advanced Data Analysis from the dropdown, and the plus icon should now appear.

Usage

You can now upload files by clicking the plus icon. Only supported file types will be shown. As a test, I uploaded a Hungarian insurance policy PDF and successfully retrieved the validity period.

There may be instances where ChatGPT can't extract information directly. In such cases, it offers alternatives like OCR, demonstrating ChatGPT's commitment to finding solutions even when initial attempts don't succeed.