Skip to main content

What‘s the Difference Between AI Smart Alerts and Custom Alerts?

C
Written by Cary Olson
Updated over 3 months ago

In OneSight, the alert system is designed to help users detect anomalies, negative content, or sudden events in real time. The system offers two alert configuration modes: AI Smart Alerts and Custom Alerts.

1. Key Differences

AI Smart Alerts

Custom Alerts

Trigger Mechanism

Automatically identifies sudden changes via built-in algorithms

Manually configure trigger conditions and thresholds

Setup Difficulty

Zero-configuration; toggle to activate

Flexible configuration; ideal for specific monitoring needs

Target Users

First-time users, those seeking "automatic monitoring"

Experienced operators and analysts familiar with the business

Supported Scenarios

Sudden volume spikes, sentiment reversals, topic surges, etc.

Mention counts, sentiment ratios, specific keywords, authors, platforms, etc.

Accuracy

Highly dependent on model judgment; suitable for overall trend monitoring

Fine-grained settings for precise business objectives

Controllability

Automatically triggered; no modifications allowed

Rules are user-defined and fully customizable

Typical Use Cases

New product launches, 24/7 basic monitoring

Crisis keyword tracking, brand campaign monitoring, sentiment intervention threshold setting, etc.

2. Selection Recommendations

  • AI Smart Alerts are ideal for early-stage deployment, especially when the monitoring focus is not yet clearly defined or team resources are limited.

  • Custom Alerts are better suited for scenarios where business needs are clearly defined and a specific response mechanism is required (e.g., daily checks, on-call handling).

  • Both alert types can be enabled simultaneously, enabling a monitoring strategy that combines broad coverage with precise focus.

If you're unsure how to configure custom alert rules or would like to use industry preset templates, feel free to contact the OneSight Customer Success Team for setup assistance and scenario-based recommendations.

Did this answer your question?