More and more companies are betting on Artificial Intelligence to streamline their processes. This week we were told that the HolaLuz company was using Atomian's Artificial Intelligence to automate the processing of judicial notifications using the Kmaleon platform, a really interesting idea in a sector that still has to modernize a lot.
AI can offer many solutions for law firms and solicitors, since a large volume of judicial notifications can be managed in a short time.
How does AI work, what exactly does it do and how?
Artificial intelligence is a very broad field that encompasses many different disciplines. Most of them are inspired by abilities of the human mind that have not yet been reproduced in computers, although there are also others that surpass those same abilities. In general, we could agree that Artificial Intelligence is the discipline that studies any new algorithm that we did not know how to build until now.
In the case of Atomian, we have built an artificial intelligence as a universal knowledge system, which learns concepts in a very similar way to how people do it. Based on this knowledge, he is able to interpret documents in their entirety to understand their meanings. It is thanks to this approach that we can interpret any text and extract knowledge about the meanings with reliability always close to 100%.
How has the AI been trained to do that job?
The construction processes of new document reading solutions with Atomian technology are very similar to the training process that would be done on a human team. Atomian needs to understand the phenomena to be described in the documents in order to make a proper interpretation. From there, you are told what the exact extraction to perform should be. Atomian's learning curves are very steep, it quickly reaches reliability levels of 95%, and then we perform various complementary processes to nominally reach 100%. These processes cover many aspects, not only those of artificial intelligence. Any project needs its dose of engineering.
What other applications does that AI have?
Many! Atomian is, as I said, a comprehensive knowledge system with the ability to understand data and texts. We could potentially build cognitive software from scratch using only Atomian, or we could let the user explain what software they need and Atomian would be able to program it itself.
How many documents can be managed per day and how?
Our current infrastructure is sized to process one million documents per day, which is our strategic goal. But we can easily make it grow, Atomian is very scalable.
You can learn more about the solution at atomian.com.