Home » Cases » AI document archiving automation: saving time with Notion

AI document archiving automation: saving time with Notion

We implemented a comprehensive solution to automate the full cycle of receiving, classifying, and archiving PDF documents. The system processes incoming documents from various channels like Gmail or Telegram bot, uses artificial intelligence (AI) to recognize and categorize them, and then integrates the data into Notion and saves the files to Google Drive. This significantly reduces manual labor, ensures high accuracy, and creates a centralized, easily accessible archive.

Document chaos and wasted time

Companies often face the problem of a large volume of incoming documents that come through various channels. Manual processing and archiving of these documents leads to:

Excessive time costs

Employees spend valuable time on the monotonous routine of manually receiving, sorting, classifying, and storing documents, which reduces their productivity and distracts them from their core responsibilities.

Data inaccuracies and search difficulties

Manual archiving increases the likelihood of errors in the classification and indexing of documents, which then makes it difficult or impossible to quickly and accurately find the information you need.

Lack of a unified, structured storage system

Without a centralized and logically organized storage system, documents can be scattered across different repositories, creating “information chaos” and impeding transparency and efficient access to data.

Difficulties with tracking and controlling documents

Insufficient automation leads to a lack of clear mechanisms for tracking the lifecycle of documents, making it difficult to monitor, audit, and ensure compliance with regulatory requirements.

Intelligent archiving automation

We developed and implemented an end-to-end solution that automates the full cycle of document management:

Step 1: Accept documents from different channels:

The system automatically accepts PDF documents sent to corporate Gmail and also allows users to send files via a special Telegram bot. The system initiates a script when a new email or Telegram message is received.

Step 3: Intelligent processing and classification:

Only PDF files are accepted. Other formats are rejected or auto-converted.

Documents undergo optical text recognition (OCR) using the Google Cloud Vision API.

Once the text is extracted, the data is sent to ChatGPT/GPT-3, which automatically generates a short summary, determines the category and keywords for each document.

Step 4: Centralized storage and structuring:

All processed data and document metadata are automatically created in the Notion DB "Document Archive" with fields: Name, Category, PL Summary, UA Summary, Keywords, GDrive Link, Status.

Files are stored in a logically organized manner in Google Drive, in folders by category and date:

Document Archive////....

Step 5: Notification and exception management system:

After archiving, the bot sends a message to Telegram informing about the category, summary, and link to the document.

Users can answer the question "Do you agree with the archiving?" with the text "Yes" or "No". If the user rejects archiving, the file is deleted from Google Drive and the corresponding record in Notion.

The system handles exceptions. For an unsupported format, the response "Format not supported, send PDF" is sent. If the OCR is less than 80%, the status is marked as "Requires Manual Review" and a notification is sent to the #archive_manual channel. If there is no category, LLM prompts the user. No duplicate files are created by MD5 checksum, the bot reports "Document is already in the archive".

Order in documents and strategic focus

The implementation of an intelligent document archiving automation system was a key step in the digital transformation of the company’s document management. This comprehensive solution not only optimized routine processes, but also provided unprecedented order and transparency in document management, allowing the business to focus on strategic management and development. Automation allows you to achieve the following measurable benefits:

Significantly reduce time and resources:

Reduce FTEs involved in the archiving process by ≥80%.