Big Cover_AI Generativa Financial Services

Generative AI for automated extraction of useful business information

A cutting-edge project based on Large Language Modelling techniques with chatGPT, to optimize response time to calls for tenders in the Financial Services sector

Generative AI

Catalyst for innovation, booster for human creativity and accelerator of automation processes. Why Generative AI makes us (already) better.

90. 110. 167.  
These are the numbers of pages of specifications that bidding offices usually have to analyze, looking for accurate and irrefutable answers to draw up the documentation needed to win the bid. 
Reducing the time it takes to research and then process the necessary information and eliminate the risk of human error can result in winning projects that are crucial to a company's business.  

This is the challenge posed by a top player in the Italian Insurance market to the Lutech Group, which has devised a Gen AI-based solution that not only makes it possible to avoid reading hundreds of pages and automate the search for specific information, but also to derive or "calculate" the punctual answer from more articulate text describing it (e.g., "5 percent of").

From searching the knowledge base to processing answers thanks to OpenAI

Underlying the solution implemented by Lutech are generative AI techniques (OpenAI): starting with the development of a reading process that first segments the text into paragraphs and then filters the information, based on NLP techniques of embedding the text, one finalizes the query by presenting it to a Large Language Model (such as GPT4) so that it processes it. 

Thus, it is possible to query very large texts, analyzing the corporate knowledge base, to process the content needed to respond to the specific call for proposals.

Large Language Modelling to process natural language and provide accurate and timely responses

The use of state-of-the-art Large Language Models, such as GPT-4, provides a powerful tool for extracting and processing information from text. These models take advantage of advanced natural language processing capabilities to understand and generate human-like text based on the input received.

In Lutech's design, leveraging a Large Language Model enables the management of a number of processes essential to automating the formulation of responses consistent with requirements:

  • Text Segmentation:
    Break down the input text into meaningful segments or paragraphs for more detailed analysis.

  • Information Extraction: 
    Use the model to extract relevant information from segmented text. This involves identifying and capturing key details. 

  • Understanding Text: 
    Take advantage of the Large Language Modelling (LLM) ability to understand and interpret the context of information. This is crucial for understanding complex language structures and articulation. 

  • Processing Questions: 
    Formulate questions based on the extracted information to seek specific details or clarification.

  • Response Generation:
    Use the LLM to generate answers to formulated questions. The model can provide human-like answers, even for questions that require reasoning or understanding of implicit information.

  • Automation: 
    Integrate these processes into a workflow to automate the entire process. This reduces the need for manual intervention and speeds up information extraction and processing activities. 

Organizations, through these models, can enhance their capabilities in handling natural language processing, making the extraction and processing of information from text more efficient, accurate and scalable.

In addition to the enhancement of search engines becoming more accurate through LLMs, the most widespread application of language models focuses on chatbots, an extremely broad field with multiple use cases. Among them:

  • Virtual Assistants for Customer Support

  • Translation

  • Content Creation

Being able to enhance Natural Language Processing (NLP) LLMs also make customer sentiment analysis more precise by evaluating customer reviews or feedback. 

The application areas of Large Language Models, with Healthcare leading the way where they can examine patient data to provide personalized treatments and optimize the diagnostic process, are set to expand further with the ability to access and analyze information and identify critical issues and areas for continuous improvement.

Contact Lutech

We invite you to read the marketing policy disclaimer.

Please enter a value
Please enter a value
Please enter a valid email address
Please enter a valid phone number
Please enter a value
Please enter a value
Please enter a value

By clicking the "Confirm" button, I declare that I have read and understood the Marketing Disclaimer

I agree to receive commercial and promotional communications relating to services and products as well as information messages relating to marketing activities, as explained in the aforementioned Disclaimer

Please select an option

An error has occurred, please try again later

Thank you for your interest!
We have received your contact request; we will be in touch shortly to further discuss your business requirements.


Digital Services for your future


Case histories