Analytics
Easily track conversation performance and the metrics that matter with NLX
What's Analytics?
Analytics provides visibility on the performance of any deployed applications in your workspace through various metrics you choose. Whether using provided templates or constructing your own, Analytics is where you set up one or more dashboards of various metrics to regularly view and monitor activity.
Analytics gives you end-to-end visibility into how your AI applications perform. This data ranges from high-level business KPIs to detailed in-flow insights on user behavior. Use dashboards to monitor trends, evaluate performance, and spot opportunities to improve your conversational experiences.
Analytics supports everything from simple metric visualization to advanced calculated formulas, external data ingestion, and in-flow node-level analysis:
Monitor app health and user behavior
Validate improvements and A/B test flow changes
Identify gaps or drop-off points in conversation design
Track business outcomes tied to conversational experiences
Trigger alerts when important metrics exceed thresholds
Analytics powers both operational oversight and strategic optimization across your AI applications.
To access, click Analytics in your workspace menu:
Key features
Build one or more dashboards to track performance over time.
Use pre-built templates or design fully customized dashboards (enterprise) with charts, filters, alerts, and external data sources.
Dashboards help you monitor metrics such as usage, completions, escalations, errors, response times, tag traversal rates, and more.
For enterprise accounts, use advanced formulas to transform raw metrics into calculated ones.
Convert units, compute ratios, combine datasets, or build entirely new KPIs.
Multi-metric panels let you compare datasets on the same chart (e.g., comparing API calls: Data requests vs. Actions invoked).
Canvas analytics shows conversation traffic and how users traverse conversation flows, directly in the canvas builder.
See how many conversations reached each node, filter by tags or dates, and identify bottlenecks or drop-off points.
This real-time inspection helps you understand user behavior and improve flow design.
Last updated