AI Agents to improve process and infrastructure observability and achieve a self-healing system for banking IT. Innovaway si prepara a strutturare casi concreti per l'integrazione dell'Agentic AI in tre ambiti strategici.
The first – he tells Antonio Burinato, General Manager of Innovaway – concerns system observability, an evolution of the current, typically siloed, process monitoring, into an end-to-end predictive capability: AI agents can automatically link performance across infrastructures and application ecosystems to processes, identifying bottlenecks and performance anomalies before they impact business continuity and customers."
The second area is supporting a“prevention and act", where AI agents enable proactive incident management.
By analyzing historical patterns, real-time metrics, and cross-system correlations, the AI can predict, for instance, potential failures in instant payment systems, blockages in KYC/AML procedures, or resource saturation during peak times, automatically triggering corrective actions before problems manifest.
The third area 'concerns Agentic AI, which can also improve operational management on unstructured open sources, processing and contextualizing information, 'for example from application logs, incident management tickets, or continually evolving documentation – to extract insights and accelerate decisions and interventions,' continues Burinato.
Innovaway's path involves solutions featuring the orchestration of multiple specialized AI agents that collaborate to complete and verify complex tasks, overcoming the limitations of a single AI model and creating more powerful, scalable, and efficient systems for even highly articulated and interconnected applications and workflows.
All of this without ever disregarding human supervision and decision-making.
"This development – continues Burinato – will start in 2026, to respond to the increasing infrastructural complexity of companies linked to the digital transformation of recent years – hybrid multi-cloud environments, constantly increasing data volumes, regulatory constraints, and greater exposure to cyber security attacks – which is also accompanied by the need for increasingly rapid response times."
"For decades, the world of IT services has been dominated by a reactive paradigm, known as 'Break-Fix' – continues Burinato. A model based on tickets, fragmented workflows, and consequent user frustration, where success was measured by the speed of resolving a problem that had already manifested. Today, this approach is not only inefficient but has become a drag on business. And the answer to this challenge comes from technology."
In fact, banks must transition from reaction to prediction and prevention, thereby anticipating and resolving possible problems before they impact the end-user, thanks to AI-driven automation.
"The truly transformative approach involves using platforms natively designed with Agentic AI at the core – emphasizes Burinato. As confirmed by analysts like Gartner and McKinsey, 'AI Agents' are the next frontier: systems capable of executing complex, multi-step tasks without direct human supervision. Applied to IT service management, Agentic AI doesn't just suggest a solution: it executes it. This allows for the creation of true 'Self-Healing' ecosystems."
AI Agents allow banks to achieve significant benefits on multiple fronts. Firstly, the optimization of operational costs (TCO) through the reduction of manual intervention, a priority aspect for finance functions.
However, the introduction of AI generates broader impacts: from a reduction in downtime to an increase in operational efficiency, thanks to the automation of routine tasks and the proactive resolution of critical issues.
"Banks can thus make a qualitative leap also on the compliance front, reaching new standards of resilience and adaptability – concludes Burinato. The IT infrastructure, in fact, becomes more robust and reactive, capable of automatically managing any critical issues without service interruptions, in line with what is required, for example, by the DORA regulation."
