AIOps

 View Only

 AIOps is becoming a decision layer, not just an operations layer

IGOR JAVIER Villota Ortiz's profile image
IGOR JAVIER Villota Ortiz posted 09/21/26 10:46 PM

As AI becomes increasingly embedded in IT operations, I think there is an architectural question worth discussing:

Should AIOps stop at predicting incidents, or should it become capable of evaluating the operational consequences of different decisions before they are executed?

In complex environments, detecting an anomaly is only the first step.

A more advanced operational architecture could combine:

Observability → AI reasoning → System modeling → Simulation → Decision → Automated action

The key difference is the introduction of a system model.

If an AIOps platform understands not only the telemetry of a system, but also its dependencies, constraints and possible states, it could potentially simulate different interventions before applying them to production.

This is where I see an interesting convergence between AIOps, Digital Twins, AI and simulation.

For example:

  • What happens if a workload is moved?
  • What is the expected impact of a capacity change?
  • How could a remediation affect downstream services?
  • Which intervention minimizes operational risk?

The question I’m exploring is therefore less about “how do we detect incidents?” and more about:

Can AIOps evolve into an intelligence layer capable of understanding, simulating and optimizing complex systems before taking action?

I'd be interested in perspectives from teams working at scale with AIOps, hybrid cloud, observability and AI-driven operations.

Especially interested in where you believe the boundary currently sits between AIOps, autonomous operations and Digital Twin-based decision systems.

Ayesha Adalat's profile image
Ayesha Adalat

Interesting perspective, Igor. I think the real shift is from “detect and alert” to “understand, reason and act.”

In complex hybrid environments, knowing that something is wrong isn't enough. The system needs context around dependencies, configurations, workloads and operational impact to determine what should happen next.

I also think the boundary between AIOps and autonomous operations will increasingly depend on how much context the AI can understand before making a decision and how safely it can validate the outcome.

The interesting question for me is: how do we move from AI recommending an action to AI understanding the potential consequences of that action before execution?