This morning’s panel at Fiber Connect Latam 2025, moderated by Mariana Rodriguez Zani, CEO of Grupo Convergencia, explored the range of impacts that AI has already shown—and promises to continue showing—across various players in the value chain. Among the common threads discussed was the growing demand for ultra-connectivity; while current AI usage remains largely centered on customer service and the pursuit of operational efficiencies. In connection with the latter, the particular context of Argentina—mostly economic—makes that pursuit of efficiencies especially crucial.
Alejandro Girardotti, Senior Director of Product, Innovation, and Strategic Alliances at Cirion, described the impact on telecommunications networks as a growing expectation of increased consumption across the two main AI processes. For model training in specific use cases, transferring information—sometimes in the form of images or videos—is required via datasets. “This drives the need for connectivity toward the central data center, and this transmission demands increasingly larger bandwidths. In the other process, inference, there’s also a need for high bandwidth between the end user and the data center—or multiple centers—as inference is distributed across edge data centers to be closer to the customer. In short, there’s a sharp increase in connectivity requirements between data centers, toward those centers, and between the user and the data center—and on top of that, we must factor in latency, which is critical for certain use cases,” he outlined.
Gerardo Renzetti, CTO of Arsat, suggested looking at AI from both of its faces. On the benefit side, it can be used in orchestrators or traffic managers, enabling more dynamic and proactive network management, and the accumulation of data allows for continuous improvement. “As a challenge, we have to consider that technological advancement forces telcos to renew or expand their infrastructure to support what AI demands. As AI access and usage increase, the demands on telcos will be massive—in terms of capacity, latency, and distributed processing,” he warned.
Miguel Zehnder, CTO of Alvis, echoed concerns about AI’s potential, especially in its generative form. “We’ll have to see what happens when large numbers of users begin using it, and how infrastructure capacity keeps up with that demand. Very good products, launched without a solid backbone behind them, often end up causing major issues. So, a key challenge is scaling AI so it can truly take off.”
From Sion, the company’s CEO, Luis Quinelli, acknowledged that there is an ongoing debate about whether traffic will actually increase. “We believe it will, due to higher content consumption—much of it increasingly generated by AI and specifically tailored to individual interests. There’s a paradigm shift happening in content creation, and in how each of us conceives content consumption,” he proposed.
Customer Service. Regarding current, concrete applications of AI, customer service is where it is most frequently observed. Sion has implemented tools in this area that allowed for automation of processes and cost reductions. Quinelli emphasized the “significant” impact on service commercialization. “This is a technology we must embrace in a context of efficiency-seeking—and within Argentina’s broader context.”
Zehnder aligned with Sion’s CEO on the efficiency gains made possible by applying AI to customer service. For the executive, in smaller companies, the use of this technology represents a chance to improve processes that previously weren’t possible. At the same time, this goes hand-in-hand with a growing need for digital resources, which will require highly qualified personnel.