How can we build intelligent, strategically designed IT operations that create tangible business value? Operations that not only react but proactively prevent issues autonomously, without human intervention? According to Viktor Weininger, CTO/COO of ABRIS, the answer lies in AIOps and NoOps, paving the way for IT to lead, not follow, business innovation.
*This arcticle was originally published in Hungarian on the pages of the online industry magazine fintechzone in June 2025.
Everyone Wants a Perfect System
Everyone wants systems that run smoothly: free of errors, fast, easy to adapt, secure, and cost-efficient to operate. Unfortunately, systems rarely meet all these expectations, especially when we’re dealing with something as complex as a core banking system. At least, not by default. But an optimal state is achievable.
How can we go beyond digital transformation and build truly intelligent, strategically designed IT operations that deliver real business value? Viktor Weininger, CTO of ABRIS, helps us navigate this journey by outlining the key steps needed to reach optimal operational performance.

Viktor Weininger, COO/CTO of ABRIS
Thanks to his decades of experience in the world of banking technology, Viktor Weininger has an in-depth understanding of how IT systems operate. As the CTO of ABRIS, he has been involved in the implementation, upgrade, and optimization of core banking systems throughout his career. In the process, he laid the foundations for a monitoring and automation framework that today supports the stable operations of more than 90 banks worldwide—impacting the lives of 250 million customers.
The State of Banking IT Operations Today: Overloaded and Reactive
Although digitalization has advanced significantly in the past decade, the operation of banking IT systems still faces considerable challenges. In traditional models, IT teams are responsible for monitoring subsystem performance, detecting and managing incidents, and ensuring smooth system processes.
The issue today is that these tasks are typically carried out manually, or at best, only partially automated.
Financial institutions often operate in heterogeneous IT environments, where software components of different types and ages must be integrated. As a result, operations teams are overloaded, fault and performance issue detection is slow, and often problems are only noticed once they’ve already impacted the business, such as in the form of customer complaints.
At the same time, bank leadership expects fast and precise answers from IT: which services are overloaded, where intervention is needed, and when development or go-live can proceed. Today’s operational models often fail to deliver this clarity.
“Our experience shows that operational tasks haven’t changed much in the past 10 years: root cause analysis still remains the biggest challenge.”
“When an incident occurs, the actual resolution teams are informed too late and lack detailed data about the surrounding context to apply workarounds. Moreover, they don’t have a complete picture of the affected business areas.”
“Unexpected slowdowns are another major challenge, especially when multiple systems are interconnected. It’s hard to localize which layer is responsible for the slowdown and determine how to fix it”, sums up the ABRIS CTO.
Most banks already use one or more monitoring and automation solutions. However, when faced with the challenges described above, it becomes clear that either the measurement points aren’t well-configured, or there’s insufficient contextual data about deviations, leaving manual methods as the only fallback.
“I believe good operations are only possible when monitoring and automation are co-located—or at the very least, fully integrated.”
“With automation, it’s not enough to just run jobs in sequence or execute simple scripts. You need to understand what’s happening and why: and that’s the role of the monitoring application. And vice versa.”
“If we detect threshold breaches during monitoring, immediate intervention is needed—and the automation tool must carry it out. Thresholds and rules also need to be continuously adjusted, automatically.”
AIOps: The Foundation of Intelligent IT Operations
The goal of AIOps (Artificial Intelligence for IT Operations) is to support system health and performance through artificial intelligence. By continuously collecting and analyzing data (such as logs, metrics, and events) AIOps can detect anomalies in real time and even forecast potential future issues. The foundation is high-quality, real-time data and meaningful statistics.
Step-by-step guide to building an AIOps platform:
1. Domain-specific, high-quality data collection
The monitoring system must thoroughly understand the application and all its integrations. This is how domain-specific knowledge is built—the essential foundation for everything else.
2. Auto-baselining and alerting
False alarms are a common problem in monitoring. A poorly configured alert threshold leads to false positives, which erode trust in the monitoring system. Yet, setting perfect thresholds is difficult, time-consuming, and requires deep expertise and they’re not static values. For instance, consider the average transaction volume on a Sunday evening versus Monday morning. These can’t possibly have the same thresholds. Based on historical data, AI can identify patterns and create dynamic limits.
3. Automated deep analysis
IT operations resemble a general health check. You go for a blood test every year. If everything’s fine, you return next year. But if a value is too high or too low, the doctor orders further tests. Likewise, we must automatically perform deeper diagnostics to get closer to the true root cause.
4. Contextual root cause analysis
Issues must be investigated comprehensively and in context. Environmental factors present at the time of an anomaly can help explain it. For example, a mobile banking slowdown might appear random until we find that one of eight app servers is running a memory-heavy process. You only spot the correlation if you examine both ends.
5. Automatic component detection
Monitoring must be complete and adaptive. When a new component enters or leaves the system, the monitoring solution should detect it automatically and extend its scope accordingly. If, for instance, a ninth app server is added, it should be monitored too.
6. Business service-level mapping
It’s crucial to establish a common language between IT and business. Define business services that describe the IT dependencies required to execute specific business functions, like opening an account or processing a card payment. These can then be monitored as units. In case of service outage, business areas and call centers can be alerted automatically.
7. Forecasting
Understand service behavior, identify patterns, and visualize them. This allows trend-based forecasting. Returning to our earlier example: if a background process consumes increasing memory on one app server, AIOps can predict resulting slowdowns or timeouts on electronic channels.
8. Automated remediation
Once the forecasted values hit predefined thresholds, the automation engine executes its response. It may terminate the rogue thread or free up memory—preventing timeouts and slowdowns.
AIOps isn’t a technology for its own sake. It’s a means to make IT operations invisible yet precise—seamlessly supporting the operations of a digital bank.
This kind of intelligent operation not only enables internal optimization but also opens the path toward the Operation-as-a-Service model.
Operation-as-a-Service: Outsourcing with Intelligence
More and more financial institutions are choosing not to manage IT operations in-house but to fully outsource them to specialized partner companies.
This is what we call the Operation-as-a-Service model. Here, the provider is responsible not only for day-to-day operations, but also for continuously monitoring, optimizing, updating, and auditing the systems.
This approach becomes truly effective when operations rest on AIOps foundations. That means the provider doesn’t just rely on human effort but uses automated tools, predefined response templates, and predictive analytics. This ensures high availability, fast incident resolution, and regulatory compliance.
From the client’s perspective, the model also reduces costs and risks. Operations are performed by a specialized partner under strict SLAs, greatly relieving internal IT and allowing teams to focus on innovation and development.
However, the real leap isn’t just outsourcing—it’s fully automating that outsourced operation. This is where the AIOps- and OaaS-based NoOps vision becomes truly visible.
NoOps: The Vision of Fully Automated Operations
NoOps (No Operations) is a vision where the entire operational workflow (from monitoring to intervention) runs without human involvement. Developers and business analysts deliver software and configurations through a well-defined CI/CD pipeline. Meanwhile, the backend system automatically scales, monitors, heals, and reports.
While full NoOps implementation is still a future goal for most large banks, the path toward it is already walkable, step by step:
- AI-based software generation and code review – AI codegen (code generation) or AI dev copilots.
- Automated deployment and rollback processes – Automated creation of test environments and alignment between development and operations.
- Infrastructure as Code – Complete environment descriptions in versioned form.
- AI-driven software testing – Automated test case generation (based on specifications/code), intelligent test data creation, change-based test prioritization (impact analysis).
- AIOps implementation – Creation of predictive and autonomous systems.
- Operation-as-a-Service – Standardized, SLA-based operations.
- Resource management – Automatically scaled resources based on patterns and forecasts through the intelligent AIOps system.
The goal isn’t to eliminate operators, developers, or testers—but to gradually make operations automated, proactive, and invisible.
This enables IT to function as a true strategic partner to the business, not just the department that gets called in during a fire drill. In the era of digitized banking, operational invisibility, reliability, and flexibility have become strategically critical.
AIOps and Operation-as-a-Service already offer practical solutions for more efficient, proactive operations. NoOps charts the future where IT doesn’t just follow business innovation but lays the groundwork for it.
Banks that begin building this future today will not only become more stable, but more competitive as well.
