Intelligence and AI: Copilot, events and reliability

Summary

Wertek AI's intelligence layer turns monitoring data into decisions: a conversational assistant (Wertek AI Copilot) that knows your assets, an Event Intelligence engine that tracks the full lifecycle of each event (onset → escalation → recovery) instead of firing isolated alarms, an operational health panel, alerts prioritized by asset criticality, and context tools: a dependency graph, field observations and a per-asset timeline.

The problem it solves

Traditional monitoring platforms deliver threshold alarms: every time a variable crosses a line, another notification goes off. The result is noise — hundreds of repeated alarms, with no way to tell a passing spike from a problem that's escalating, and never confirming whether the problem was resolved. At the same time, the team's knowledge (what happened last time, what depends on what, what was observed in the field) lives in people's heads. When someone asks "how's my plant doing today?", no one can answer with data. More context in market pain points.

What it does

Wertek AI Copilot: a conversational assistant built into the platform. It answers questions about your assets, measurements and maintenance in your language, with your organization's context — not generic answers. Its answers are grounded in your real data: it can cite the complete history per failure type (when the last one happened, how long it lasted, how many times it has occurred), refers to assets by name (not technical identifiers), distinguishes which device detected each event and which asset it is about, and shows cards for the asset in question — drives included. It is honest about its limits: if a data point is outside the queried window it says so, rather than claiming the event never happened. From the same answer it can suggest actions: create a work order or an automation rule.
Event Intelligence: every anomaly is handled as an incident with a lifecycle: the onset is detected, its escalation is tracked, and it only closes when the equipment's own data confirms recovery. Closing a work order does not close the incident: the resolution is verified with the measurement. Every incident lands on the asset the condition is about — the meter detects it, the panel owns it — with the provenance visible (which instrument detected it).
Operational Health: a reliability view by plant, area and criticality — what proportion of incidents get resolved, where they concentrate, and how the operation evolves.
Risk-based alert prioritization: the alert center ranks by asset criticality and severity, and groups repeated episodes to eliminate noise — an intermittent problem generates one episode, not thousands of notifications.
Plant graph: a live, interactive map of the assets and their declared connections (what measures, drives, and feeds what), with a Constellation view, assets with an active condition that pulse, per-connection-type lenses, and an impact radius to understand what is affected if a piece of equipment fails.
Field observations: structured capture of what the technician sees on site — photo, title, description, severity, tags and location — linked to the asset and visible to the whole team.
Asset timeline: a unified chronological history of measurements, events, orders and observations per asset.
Studio / Engineering Lab BETA: an advanced analysis space for engineering users, available to beta testers.
ISO 55001-style reliability module (risk register, maturity, audits) ROADMAP.

Benefits

Less noise, more signal: lifecycle-based episodes instead of repeated alarms; the alert queue holds real problems.
Verified resolution: you know from the data whether the problem was actually fixed, not just whether someone "closed the ticket."
Defensible prioritization: what's critical comes first, according to the criticality your team assigned to each asset.
Knowledge stops living in people's heads: observations, history and dependencies stay in the platform, queryable by anyone — even by asking the Copilot.

Experience on the web platform

In the sidebar, the Intelligence group: Event Intelligence (incidents and their lifecycle), Operational Health (reliability), Automations, Wertek AI Copilot (chat) and Studio BETA. In the Diagnostics group live Timeline, Observations and Graph. The prioritized Alerts are in the overview group, next to the main Dashboard.

Experience on the mobile app

From the phone: a unified alert center, an incidents list with filters by plant/area/criticality, an operational health and reliability panel, and capture of observations with a photo directly on the asset (even by scanning its QR/NFC on site). The conversational Copilot is currently a web-platform experience.

Availability by plan

Alerts, Event Intelligence and operational health are available from the energy-monitoring plans onward. Timeline and observations require the Technical plan or higher; the Copilot and automations, the Monitoring plan or higher; the dependency graph, the Prescriptive plan. Studio is limited to beta testers. Amounts and detail in Plans and pricing.

Condition diagnostics and monitoring — the source of the data this layer interprets.
Energy and power quality — energy incidents with the same lifecycle.
CMMS and field service — turn incidents into work orders with evidence.
Platform overview