At Neuron Solutions, we have long been using Large Language Models (LLMs) to create intuitive chatbots and streamline process automations. Traditional approaches offered immense benefits—enabling faster customer support, refined data insights, and the ability to execute various everyday tasks more efficiently. However, as business challenges grow in complexity, we realized something more advanced was needed. Enter agents—a new breed of AI powered by LLMs, yet designed to think, plan, and act autonomously across a wide range of interconnected tasks. An agent can be defined as an autonomous AI entity that not only processes language but is also capable of orchestrating and executing multi-step plans. If we want to give a definiton, “an agent can perceive its environment, take actions autonomously

The following summary is based on the March 17, 2025, discussion on Millásreggeli, featuring Levente Szabados, associate professor at the Frankfurt School and co-founder of Neuron Solutions. Szabados shared his insights on the future of programming automation and its potential impact on the industry.   Programming as an automatable activity The evolution of programming languages presents an intriguing paradox: while numerous new languages have emerged since the late 1960s, the fundamental process of programming has remained relatively unchanged for decades. Szabados highlighted a well-known saying among developers: “If something takes 30 minutes to complete manually, it is worth spending an hour and a half automating it.” This mindset is now being applied to programming itself, as artificial intelligence (AI) becomes

Analyzing the electricity consumption of a hypermarket at quarter-hour intervals is a serious challenge in itself, but what happens if we want to increase the accuracy of the forecast? The study by Csilla Obádovics and Levente Szabados seeks to answer exactly this question, presenting a methodological experiment that focuses on the comparison of traditional forecasting methods and state-of-the-art artificial intelligence-based models.  Background of the research  The authors examined the detailed consumption data of a Hungarian hypermarket between September 2007 and July 2008, which were recorded in quarter-hour breakdowns. The aim was to forecast the electricity consumption of industrial consumers for 48 hours in advance, taking into account weather data, the number of visitors and other variables. One of the main