blog

A look back at the CRiP x ServicePilot AIOps workshop

A look back at the <span class='blue'>CRiP</span> x ServicePilot AIOps workshop
November 19th, 2026

How do AIOps and silo-free observability improve IT?

On November 18, 2026, the CRiP (Infrastructure and Production Managers Club) hosted a strategic workshop with ServicePilot, focused on the challenges of modern IT operations. This session addressed a topic at the heart of IT teams’ concerns: "AIOps + Silo-Free Observability: From Signal to Action".

Logo CRiP

As information systems become increasingly distributed, hybrid and dynamic (cloud, Kubernetes, microservices), traditional approaches are reaching their limits. Here are the key takeaways from this discussion, centered on five major technological pillars.

Breaking down silos to understand operational reality

Relying on a growing number of specialized monitoring tools is no longer enough. The performance of a digital service is never limited to the health of a single component in isolation. To move beyond simple technical monitoring to a true understanding of the operational situation, it is essential to consolidate infrastructure monitoring, application observability and knowledge of dependencies.

At the same time, OpenTelemetry has established itself as an essential foundation: by standardizing the collection and transport of signals (metrics, logs, traces) across heterogeneous environments, it ensures reliable, contextualized and directly actionable data.

5 key topics at the heart of the discussions

During this workshop, the speakers broke down the AIOps value chain across five specific technical areas:

1. The dynamic topology of layers 2 through 7

Understanding an incident requires the ability to instantly visualize the interactions between the network, the infrastructure and the application layers. Automated end-to-end mapping links each physical or virtual component to the business service it supports.

2. Anomaly detection via Machine Learning

Static thresholds generate too many false positives. Using Machine Learning, the platform learns the normal behavior of each system and automatically identifies significant deviations, even in complex environments where certain variations are common.

3. Root Cause Analysis (RCA)

Detecting a weak signal is only the first step. By correlating events, metrics, traces and topology in real time, the analysis engine traces the chain of causality to quickly isolate the likely source of a multi-layer outage, helping reduce the mean time to resolution (MTTR).

4. Generative AI-assisted diagnosis

To leverage this wealth of data without adding cognitive complexity for DevOps and Ops teams, generative AI acts as a co-pilot. It summarizes incidents, explains correlations in natural language, draws on the existing knowledge base and suggests remediation paths.

5. Integrating ServicePilot as a MCP server

One of the highlights of the workshop focused on the openness and interoperability of AI models. Using a MCP (Model Context Protocol) server enables AI assistants to be securely connected to real-time observability data and operational tools using a standardized API. This interface transforms diagnosis into concrete, context-aware actions for AI agents.

Artificial Intelligence at the service of human expertise

The message conveyed throughout this workshop is clear: the goal of AIOps is not to replace human expertise, but to augment it. By filtering out operational noise to deliver the right signal, the right context and the right information at the right time, IT teams can focus on continuous improvement and business value.

From standardized data collection to generative AI-powered assistants, unified observability is setting the standard today for the IT operations of tomorrow.

Did you like the article? Feel free to share it