Building specialized AI agent teams that work side by side with people inside a company, without giving up the governance that European regulators require. Data stays in Europe, and every activity is fully traceable.
That's the vision Exelab brought to the stage at Agent ORG on May 26, 2026, at the Terrazza Palestro in Milan. They didn't just talk about it; they ran a live simulation showing twenty-seven people working alongside eleven specialized AI agents, in front of an audience of Italian managers and decision-makers.
This wasn't a demo or a prototype. It was a production-ready organization: people and AI agents collaborating, coordinating, and working together as a team toward shared business goals.
To close the evening, Emanuele Caronia, CEO of Exelab, announced the launch of Polyant, the platform Exelab built to make all of this possible.
"Applying AI to existing processes isn't enough. You have to transform those processes and let them evolve. That's the only way this technology can have a truly meaningful impact."
- Emanuele Caronia, CEO and founder of Exelab
Polyant is a platform designed to build teams of specialized AI agents that collaborate with different teams across a business.
Each agent has a defined role, a specific set of tools, and declared governance rules. Agents communicate with each other, share context and tasks, and return processed outputs ready to support decision-making, creating a synergistic, complementary collaboration between people and AI.
The platform is built to work across very different industries: banking and financial services, insurance, healthcare, telecoms, manufacturing, and more broadly any context where governance, traceability, data protection, and regulatory compliance are essential requirements for putting AI agents into production.
For companies that need support along the way, Exelab offers tailored programs, from defining the first agent all the way through to deploying full agent teams with diverse roles and responsibilities.
Polyant is built around five characteristics that set it apart from other agentic platforms available today: it is European by design, with GDPR compliance and data that stays within the European Union; it is sovereign, because it can run on-premise, on the customer's own cloud, or on infrastructure managed by Exelab, tailored to specific needs; it is open, released under the AGPLv3 open source license with public code on GitHub; it is governed, with ten built-in controls that kick in before, during, and after every agent decision, plus an audit trail and logs queryable via SQL; and finally it is operational, with native multi-channel support for Telegram, Slack, and WhatsApp, and compatibility with Anthropic, OpenAI, and Mistral models, the full AWS Bedrock model catalog, and internally hosted open source models.
The EU AI Act will become fully applicable to high-risk systems in August 2026, and DORA is already in force for banks and insurers. Companies deploying AI agents in production without a governance framework are essentially building on regulatory quicksand.
Polyant was created to address exactly this need, meeting customers with flexibility. From fully self-directed adoption starting from the open source code, to guided adoption programs led by Exelab for Italian and European companies that want to bring AI agents into production in a governed, sustainable way.
Learn more at polyant.ai