Dashboards

Track simulation performance and compare results across runs

Simulation dashboards give you a structured view of how your agent performs across test runs. Use them to validate quality before releases, verify new features, and track whether changes improve or regress agent behavior over time.

What Simulation Dashboards Show

Each simulation dashboard aggregates data from your test runs and surfaces the signals that matter for pre-production confidence:

• Pass/fail rates — the percentage of conversations where your agent met its defined goals
• Custom Metric trends — scores for every metric you’ve configured, tracked across runs
• Hallucination and redundancy flags — built-in checks surfaced per run
• Latency distribution — response time patterns across all calls in a run
• Call status breakdown — successful, escalated, dropped, or failed conversations

Common Use Cases

Pre-Release Quality Gates

Build a dashboard that tracks your core metrics across simulation runs. Before each release, run your simulation suite and check whether the dashboard shows scores at or above your quality threshold. If scores dip, investigate the failing calls before shipping.

Feature Verification

When you add a new capability to your agent — a new tool, a different prompt strategy, or an updated knowledge base — create a simulation that exercises that feature specifically. The dashboard lets you see whether the feature performs as intended across a range of Digital Human scenarios.

Regression Detection

Run the same simulation repeatedly after changes to your agent’s configuration. The dashboard’s trend view shows whether scores are stable, improving, or regressing. Pair this with GitHub Actions to automate regression checks in CI.

Structuring Your Dashboards

Strategy	When to use
One dashboard per agent	Small teams with a single agent and a manageable number of simulations
One dashboard per feature	Agents with distinct capabilities that need independent quality tracking
One dashboard per release cycle	Teams that run simulation suites as part of a release checklist

Keep simulation dashboards focused. A dashboard that tracks too many unrelated metrics becomes noisy. Group related simulations and metrics together so each dashboard tells a clear story.

Next Steps

Simulation Results

Understand the metrics and call-level data behind each run.

Custom Metrics

Define the evaluation criteria your dashboards track.

GitHub Actions

Automate simulation runs and quality gates in CI.

Dashboards Overview

Learn about both simulation and observability dashboard surfaces.