What you are seeing in the Eyer interactive demo

The Eyer interactive demo at sim.eyer.ai shows real industrial time-series data with Eyer's analysis applied. It is not a synthetic scenario constructed to look impressive. It is an anonymised dataset from an actual industrial operation, with the three-baseline model, correlation analysis, and alert layer built on top.

Here is a guide to reading what you are seeing.

The three baseline bands

The most visible feature of the demo is the three shaded bands that appear around the historical time-series data for each metric. These are the three dynamic baselines Eyer learns from the historical data: the lower baseline, the central baseline, and the upper baseline.

The bands are not fixed. They widen and narrow as the operating context changes — during high-load periods, during transitions, during expected process cycles. A metric tracking inside its normal band is behaving as Eyer expects based on its operational history. A metric tracking outside its band is deviating from learned normal.

Notice that the deviation Eyer flags often appears well before the line at which a traditional alarm would fire. That gap — between the fingerprint deviation and the threshold crossing — is the early detection window.

The correlation panel

When you select a flagged deviation in the demo, the correlation panel shows which other metrics are deviating in a statistically correlated pattern. This is the cascade view: the primary signal (the first to deviate from its baseline) and the correlated signals (those that deviate in a pattern consistent with their historical relationship to the primary signal).

In a traditional alarm system, each of these signals would generate a separate alarm event. Eyer presents them as a single event: one root cause, one set of correlated consequences, one alert that carries the full cascade context.

The alert detail

The alert detail panel shows the decision-ready briefing that Eyer generates from each flagged event: what is deviating, how far from baseline and in which direction, which signals are correlated, what the likely cause is based on the historical pattern, and what action is recommended.

This is not a number and a red flag. It is the operational context an engineer needs to make a decision without additional investigation. The alert is designed to be actionable on first reading.

What the demo does not show

The demo shows Eyer's analysis on a single anonymised dataset. It does not show the full range of what Eyer builds for a specific operation: the complete fingerprint across every metric, the full correlation map, the continuous adaptation of baselines as the operation evolves, or the agentic AI layer that the MCP server enables.

The fastest way to see all of those is to run Eyer on your own historical data via the Fast Forward. One week from data receipt to POV report.