Vicomtech and Tubacex collaborate in the MERLIN project to automate the extraction of content from their RFPs (Request for Proposal).

A Request for Proposals (RFP) is a type of commercial document that advertises a project, describes it and solicits bids from qualified contractors to complete it.

In the case of Tubacex Tubos Inoxidables (TTI), a company that provides tubular solutions in large projects, bases its contracting model on exhaustively analyzing this type of solutions and giving them a commercially and technically viable response.

Currently, most companies in the industry, including Tubacex, devote a great deal of effort and human resources to the task of analyzing RFPs, as they require their commercial and technical staff to read the documents directly, identify and filter critical information, collect all key points and review the request on a recurring basis during the development of a response.

As this is considered a very time and resource-intensive process that can lead to potential human error, an opportunity for substantial improvement has been identified with a collaboration between Vicomtech and Tubacex TTI.

 MERLIN, a project that digitizes the process of extracting relevant information from Tubacex's RFPs.

The MERLIN project was born from the need to optimize and filter the extraction of information from RFPs through the digitization of the procedure thanks to Artificial Intelligence technologies, more specifically Generative Models. And this is where Vicomtech comes in. 

Advances in Deep Learning models have significantly boosted Natural Language Processing (NLP), leading to the creation of the popular Large Language Models (LLMs). These have been transformational in the field thanks to their Generative Models, which allow to generate text in a coherent way, understand context more accurately and offer more advanced solutions in multiple applications.

Vicomtech is able, through generative models, to significantly improve the accuracy of RFP analysis, reducing the manual workload and increasing the efficiency of proposal management. This results in greater automation, minimizing errors and optimizing response time in the bidding process.

The MERLIN project has a duration of 3 years and is part of the Basque Government's Hazitek 2024 grant program.