Artificial agents, a governance architecture is needed

"In Italy, by 2024, investments in artificial intelligence will have risen to approximately €1.2 billion, out of a total of €40 billion in the information and communications technology market as a whole. We are at the beginning of a story, a revolution that is rapidly being implemented in companies and will yield results in the medium term. Patience and investment are needed, but one thing is certain: the cost of building agentic AI architectures is lower than the cost of implementing an IT architecture." The analysis by Stefano Sperimborgo, Data & AI Lead for Italy, Greece, and Central Europe at Accenture, is clear and sheds light on one of the many issues accompanying the spread of new AI technologies in Italy. Which one? The failure associated with the adoption of these tools. The risk exists, but those who have already been "burned" in this regard (experiencing premature project failures) have underestimated the management of AI consumption and the resources required for model maintenance and training.

"The agent governance architecture is key to winning the AI challenge," Sperimborgo explained to Il Sole 24 Ore, "because it is this architecture that allows us to control costs, organize and optimize data, and choose Gen AI models—from large-format to so-called Small Language—optimized for real needs and specific applications. Agents offer an opportunity to rethink processes and make them more efficient; they reimagine work and the way we do business."

The agentic revolution, in short, promises benefits but must be guided by specific requirements, which are, in fact, those underlying AI Refinery, the project that Accenture launched in Europe in close collaboration with Nvidia. Since its launch (last October), the company has filed 55 patent applications in 10 countries and has engaged companies such as the German group Kion AG, the beauty startup Noli (with support from L'Oréal), and multinationals such as Nestlé and Unilever. Last March, it introduced a tool (Agent Builder) that allows agents to be easily customized in code-free environments, and the stated goal is to develop 100 industry-specific agentic solutions by the end of this year. The essence of this project? Helping large European companies launch customized adoption journeys through AI agent systems (the first multi-system collaboration solutions between AI agents are already available), maintaining full control over critical data and fostering the adoption of AI-based robotics in the most strategic sectors, benefiting warehouse, logistics, and production processes.

"AI Refinery," Sperimborgo explained, "is a central hub for creating and managing agents, overcoming the limitations of a forced choice between one cloud architecture and another and becoming a key component of tomorrow's hybrid architecture, which will combine IT components and AI tools. At the same time, it is also an example of an agnostic approach to technology that moves toward digital sovereignty and infrastructures compliant with the dictates of the Digital Act and the AI Act." Nvidia's engineering expertise was crucial in developing an architecture capable of managing costs, managing LLM model development frameworks like FrugalGPT, and preparing the knowledge base (data) for AI applications, vectorizing documents and organizing agents in a hierarchical manner. "Agents," the Accenture manager added, "are the new workforce to be entrusted with operational tasks, alongside humans, who are their supervisors. The refinery should therefore be seen as an accelerator for the construction of agents for the various vertical sectors, providing advanced model optimization tools and empowering companies to use multiple tools in an orchestrated manner, so as not to lose control of costs and consumption.

Italy, on this journey, is at the beginning of the scale, in line with the rest of Europe but far from the adoption levels of the US. No one, as Sperimborgo finally observed, has yet implemented an agent architecture, but it is certain that this is now a priority for many and that large companies are much more aware of its importance and necessity. The acceleration seen in the last three to four months, with projects expected to be implemented in the next six to twelve months, seems promising.