When Diego Tuccillo and his partners launched Deduce Data Solutions at the end of 2020, artificial intelligence was not yet a media focal point or a priority on the agendas of large companies. ChatGPT was unknown to the general public, and the current enthusiasm for AI was barely hinted at. Betting on scientifically-based industrial AI before its popularization is key to understanding the development of this Cantabrian firm. "We are children of the pandemic," summarizes Tuccillo, CEO of the company. Deduce Data Solutions was born as a spinoff from the University of Cantabria. Its four founders—Diego Tuccillo (CEO), Biuse Casaponsa (COO), Steven Van Vaerenbergh, and Carlos Blanco—are researchers with over ten years of experience in machine learning, data mining, and predictive analysis, having published more than a hundred scientific articles in specialized journals. This scientific background is the project's differential pillar. "We wanted to bring the rigor of research to the private sector," explains Tuccillo. "Many companies invest in AI without achieving the expected return because they do not apply the right criteria from the start." Deduce's proposal focuses on explainable AI: systems capable of justifying their decisions in a way that is understandable to operators, technicians, or executives. In industrial environments, where traceability and safety are essential, entrusting critical tasks to opaque algorithms is not viable. "Companies need to understand how decisions are made," emphasizes Tuccillo. This requirement is even greater in sectors like legal, where Deduce automates regulatory document flows: "In that field, the system cannot invent information. Our model does not hallucinate, does not make anything up, and guarantees total security." To achieve this, Deduce develops its solutions from scratch, using languages like Python and TensorFlow, without relying on external services that may change their conditions or be interrupted. This technological independence allows deploying tools on the client's infrastructure, ensuring that data never leaves their servers and remains completely secure. Additionally, total control over development facilitates adaptation to specific cases and the incorporation of new functionalities, making their models scalable. In a context where digital sovereignty is a priority for governments and large companies, this approach is an added value. Deduce's catalog includes customized solutions and specialized proprietary products. Among them are Hefesto, which optimizes energy consumption in steel plants; Methis, focused on operational planning in foundries; Pythia, which analyzes and prescribes in automotive plants by crossing variables in real-time; Calipso-AI, which applies computer vision and logistical simulation in ports; Cassandra-IA, a conversational layer for querying industrial data in natural language; and Doxia, a document manager for industrial and legal environments that can operate offline. All of Deduce's solutions share the goal of improving energy efficiency and optimizing processes, especially in sectors where electricity costs are a key competitiveness factor. "We achieve cost reductions of between 10% and 20% for our clients," notes Tuccillo. In facilities with high energy consumption, these percentages represent significant savings. The team has grown from the four founders to 16 professionals. In four years, they have executed more than 24 projects for clients such as Nissan, Seat, Bridgestone, Siemens, Enel, Reinosa Forging & Casting, Iberdrola, and two municipalities. Deduce has been recognized for three consecutive years (2024-2026) as one of the top hundred startups in Spanish technology parks by the APTE and received the award for innovative digitalization company from the Cantabria Chamber in 2025. Additionally, it has been in Nvidia's Inception acceleration program for three years. Thanks to its trajectory, Sodercan has supported the development of Demeter-AI, a tool for intelligent soil moisture management in the agri-food sector. Tuccillo identifies both potential and barriers in the adoption of AI in the market. "There is a lot of enthusiasm, but structure is lacking. Many initiatives remain in the pilot phase and do not advance." To overcome this challenge, Deduce applies a three-phase methodology—diagnosis, development, and deployment—that prioritizes rigorous analysis of each case before any development. This commitment to quality has established the company as a technological benchmark with great growth potential. Deduce's goal for 2026 is to scale its products in the Spanish and European markets. The search for financial partners is on its roadmap, although Tuccillo is clear: "We come from science and research. Our priority is to create a truly differential product. When that is evident, the rest comes as a consequence of the value we provide." Source: eldiariomontanes.es