Principled Intelligence , an Italian AI deeptech start-up specializing in technologies for the control and governance of artificial intelligence, announces the closing of a €1.85 million pre-seed round.
The round was led by the National Technology Transfer Hub for Artificial Intelligence and Cybersecurity, established by CDP Venture Capital in partnership with Scientifica Venture Capital, together with BlackSheep, a venture capital fund focused on artificial intelligence investments, with the participation of Eden Ventures.
The transaction will enable the company to strengthen its team and accelerate the development of its enterprise SaaS platform for AI control and governance, an infrastructure designed to allow organizations to safely adopt and integrate the latest technologies, including generative AI and large language models, even in highly regulated environments.
For every company, the question is no longer whether to adopt AI, but how to do so responsibly. Companies that are able to control and govern this technology will gain an immediate competitive advantage, becoming benchmarks in their respective industries.
However, the adoption of AI in products, services and business processes is often undermined by its unpredictability. On the one hand, these systems are now capable of delivering highly sophisticated solutions to complex problems; on the other, their high degree of autonomy can lead to unexpected behaviors, with serious consequences for a company’s reputation and financial performance, ultimately eroding the trust of customers and stakeholders.
Principled Intelligence was founded with the ambition to address this challenge: to govern behavior and control the decisions of AI integrated into enterprise systems and products in a simple, transparent way, aligned with the unique principles of each organization.
To bridge this gap, Principled Intelligence is developing an infrastructural control and governance layer for AI, designed to be easily integrated into existing systems. The platform enables organizations to define their own operating principles, translate them into verifiable criteria, simulate realistic interactions and continuously monitor AI behavior, significantly reducing risk while ensuring controlled adaptability, consistency and compliance with both internal principles and regulatory requirements.
Unlike existing AI governance solutions, Principled Intelligence’s platform stands out for its ability to be easily customized through natural language and for providing in-depth explanations in the event of violations, making governance a clear, inspectable and controllable process even for non-technical users.
The technological core of Principled Intelligence is built around small language models specialized in governance, compliance, guardrailing and oversight. These models are engineered to understand an organization’s defined principles, apply them in real time and operate on-premise on standard enterprise servers, ensuring data confidentiality and compliance with privacy requirements.
In early pilot deployments, Principled Intelligence’s models have demonstrated higher accuracy, speed and efficiency than state-of-the-art approaches based on large language models, enabling more reliable and economically sustainable AI control in enterprise environments.
Principled Intelligence was founded by Simone Conia and Edoardo Barba, building on years of academic research in Artificial Intelligence and Natural Language Processing at Sapienza University of Rome. During his time at Apple in the United States, Simone contributed to research on large language models aimed at improving the quality, reliability and factual accuracy of conversational systems. After returning to Italy, Simone resumed his collaboration with Edoardo and together they led the technical development of Minerva LLM, the first family of large language models trained from scratch on Italian-language data. Today, Minerva LLM is adopted by universities, companies and developers, with more than 300,000 downloads.
The team thus brings together cutting-edge research, innovation in multilingual AI systems and hands-on experience in the international enterprise environment.
Their mission is to put AI control back into the hands of organizations, in a simple and transparent way, enabling companies to adopt AI with confidence even across their most critical processes.
“In the coming years, trust – not just capabilities – will determine the success of artificial intelligence systems,” says Simone Conia, Co-founder and CEO of Principled Intelligence.“Enterprises have powerful models and extremely valuable data, yet they often hesitate to deploy them due to the reputational, financial and liability risks involved. What’s missing is a clear and transparent way to verify that AI truly behaves in line with their principles, policies and brand identity. Principled Intelligence was created to fill this gap: we help organizations guide AI using their own principles, making it trustworthy, controllable and therefore truly scalable.”
Umberto Bottesini, Managing Partner at BlackSheep Fund, adds “Companies have recognized how indispensable AI has become for evolving operations and business models, but they have also realized that without an adequate level of control, this technology cannot be applied to the processes that matter most. Models are becoming increasingly powerful, adoption is accelerating, yet the ability to govern them is not keeping pace. For these reasons, we chose to invest in Principled Intelligence. The team has developed a platform with the potential to become the European standard for AI control, enabling companies to deploy advanced models without exposing themselves to reputational or regulatory risks and turning governance into a tangible competitive advantage.”
“Artificial intelligence is now an integral part of our everyday work, opening up unprecedented scenarios while bringing increasingly complex challenges in terms of information security and data privacy. This is why we place great trust in the Principled Intelligence team: the future will be shaped by companies’ ability to govern AI-based tools in a transparent and secure way and we are confident that, thanks to their technological expertise and vision, they will establish themselves as leaders in this field,” said Claudia Pingue, Head of the Tech Transfer Fund at CDP Venture Capital.
“AI is increasingly becoming embedded in companies’ core processes, but it requires robust governance mechanisms to reduce risk, enhance security, and make the competitive advantage generated by its outputs truly reliable. This is why we chose to invest in Principled Intelligence and in its ability to make all of this possible,” said Riccardo D’Alessandri, Managing Partner at Scientifica Venture Capital.
