LIS Data Solutions – GPT Up
LIS Data Solutions
Sector: Generalist
Business Case
We have developed a system that allows us to ask it about the information we upload based on LLM. For example, we have a specific one for legal questions, for factory assistance to maintenance staff, for motorcycle maintenance, for transport prices, to consult a company’s sales cross-checked against the databases or to ask about a company’s generic processes.
Objectives
The main aim is for it to be an assistant that helps employees to improve and simplify their day-to-day work, saving them time and optimising their performance.
Use case
By means of natural language, it interprets the question and, in a given context, searches for the answers and provides them by means of a language, which is also natural. This system, unlike ChatGPT, does not invent answers. If it does not know the answer, it notifies you.
Infrastructure
Cloud
Technology
Automatic or Deep Learning AI technologies that generate written or spoken language, images or videos Text mining
Data
Data on the manufacturing processes of the metal parts: machine used, parameters of the pressing machine, part being manufactured, place where breaks occur, physical-chemical characteristics of the materials used, etc.
Resources
The development has been conducted 100% in-house. We have assigned some 8 people to it, R&D staff, who have been researching this solution for more than 10 months. It is currently an LIs product that is marketed in an initial Setup format plus a monthly recurrence. The end customer is required to undergo a three-month period for the initial set-up and the implementation is completed with training.
Difficulties and learning
Of all kinds, mainly the processing of text files and how to process them efficiently. For example, the Spanish Official State Gazette (BOE) is processed article by article while a plain text is processed by titles or pages.
KPIs (business impact and metrics of the model)
The ROI of the project is demonstrated in 12 months.
Funding
No
Collaborators, Partners
No
