Digital transformation in Legal Auditing processes

Transforming one of the auditing leaders in LatAm thanks to the power of AI.

About GCG Control

GCG Control specializes in document control and contractor auditing across Latin America. They protect companies by identifying risks before they become issues, offering tailored solutions through their proprietary technology. Their services help reduce union claims, labor lawsuits, entry delays, and overpaid hours, ensuring compliance and operational efficiency for their clients.

Human Control since the company’s inception

GCG is a company whose objective is to analyze the validity of different legal documents for its end clients. This validation is carried out heuristically through manual processes, with the document having to analyze, review, and approve a certain number of validations. This manual process becomes a task that takes a lot of effort and time due to the number of files that GCG has to process. In addition, there is also the complexity that each type of file can come in different forms, and even how these are structured can be undefined per document, which is why the process of validating them can mean a great effort.
Any new big customer or any new document type to revise needs to scale in company personnel, becoming a linear growth in payroll for each time the business needs to address a new challenge.

A digital strategic pivot in the legal ecosystem

Martin Mora, the company CEO, asked DinoCloud for help in their digital transformation journey. Since the company didn’t find any online solution for Latin American countries’ legal documentation analysis, there wasn’t an out-of-the-box alternative for human analysis and intervention in their actual business processes.

Martin explained the need to grow exponentially without the need to increase personnel each single time and the need to create a digital solution that could open doors to new law firms in the pursuit of online legal processing software.


The challenge: Address several different document formats

Building a digital solution for GCG Control involves addressing the complexity and volume of diverse legal documents that need validation. The current manual process is labor-intensive and unsustainable for scaling, as each new client or document type requires additional personnel. The challenge is to develop a digital system that can efficiently analyze, review, and validate these documents, ensuring legal compliance without the need for proportional increases in staff. This system must handle various document formats and structures, automate the validation process, and support exponential business growth.


Proposed Analysis

To work in this particular case, Dinocloud decided to analyze different AI techniques, particularly natural language processing (NLP) approaches. As the files can come in different formats (PDF, PNG, JPEG), it was decided to extract the text from the documents using OCR (Optical Character Recognition) techniques. For image identifications (such as signatures, logos, QR detection, etc.) it was proposed to use a classification technique using convolutional neural networks. Then, for the recognition of information from the subsequently extracted text, named entity recognition (NER) techniques will be used, relying on current large language models.

From the extraction of information, validations could be carried out through configurations according to the type of file in question, accelerating its implementation.

Project result

  • Automating the legal document validation process: The implementation of the AI ​​solution completely automated the legal document validation process, eliminating the need for manual review. This resulted in a significant reduction in the time and effort required to process documents.
  • Improved validation accuracy: The AI ​​solution uses cutting-edge techniques to extract and analyze information from documents, ensuring greater validation accuracy compared to manual methods.
  • Cost reduction: Process automation and improved validation accuracy resulted in significant cost reduction for GCG.
  • Improved operational efficiency: Automating the process allowed GCG to improve its operational efficiency, freeing up time and resources for employees to focus on higher value-added tasks.
  • Greater scalability: The AI ​​solution is scalable, meaning GCG can process a higher volume of documents without needing to increase staff or infrastructure.

GenAI Business Impact for GCG Control

The implementation of the AI-driven digital solution significantly improved GCG Control’s business KPIs. By automating the legal document validation process, the company reduced processing time by 99%, allowing it to handle a larger volume of documents without increasing staff. Usually one of GCG collaborators takes between 1 to 5 minutes to process a single document verification, while the solution created lasts between 500 and 1500 ms to do so.

This automation also led to a 50% reduction in operational costs by eliminating the need for manual reviews. GCG Control still keeps its team of 6 collaborators reviewing documents. They’re focusing on the documents the solution still doesn’t process. They will be relying on documents that the software doesn’t process until they can process everything on their own. One of them is acting as a QA validator in case there is a false positive in the algorithm. With the cost structure GCG Control has, they were able to process between 14,000 and 15,000 documents per month. Now they can process more than 1,000,000 documents per month on classic structures and dedicate a buffer of 14,000-15,000 on specialized documents and QA.

Additionally, validation accuracy improved by 30%, minimizing errors and ensuring compliance with local regulations. The algorithm threw an error rate of 1% in the last quarter, while human errors were about 5%.  

Overall, the enhanced efficiency and scalability enabled GCG to focus on high-value tasks and support business growth effectively.

Looking for your Ideal Tech Partner?

We’d love to lend you a hand.

Get in touch

(*) Required fields