Most IT monitoring systems tell engineers when something has already gone wrong.
Startup Empirik is taking a different approach: using AI to identify changes that could cause infrastructure problems before an outage occurs.
The Sequoia-incubated company has spun out independently with $21 million in seed funding from Sequoia, Canapi and Alumni Ventures. (TechCrunch)
Moving from monitoring to prediction
Modern companies can operate thousands of interconnected systems.
A seemingly harmless software update can affect another application, database or service somewhere else in the infrastructure.
That makes it increasingly difficult for engineers to understand every dependency.
Empirik tracks system changes and attempts to infer how those changes could affect the wider environment. (TechCrunch)
The infrastructure problem
Traditional observability tools are generally designed to help teams understand what is happening inside a system.
Empirik is focused on another question:
What could happen next?
The company’s system is intended to function like an automated traffic controller for infrastructure changes—allowing lower-risk changes through while flagging more complicated or potentially dangerous updates for human review. (TechCrunch)
Why AI fits the problem
The complexity of modern IT infrastructure makes this type of analysis difficult for humans alone.
There can be thousands of dependencies between services.
AI systems can potentially analyze large numbers of relationships simultaneously and identify patterns that might otherwise be difficult to spot.
That doesn’t eliminate the need for engineers. Instead, it can provide them with another layer of analysis before changes reach production.
A new generation of infrastructure startups
Empirik illustrates a broader startup trend: applying AI to the systems that already run the digital economy.
Rather than building another consumer-facing AI application, startups are increasingly targeting invisible infrastructure problems.
These businesses may not generate the same public attention as AI assistants, but they address challenges that become more important as companies operate increasingly complex software environments.
If AI can help prevent outages rather than merely diagnose them afterward, it could change how enterprise infrastructure teams approach reliability.