DemoBitesDemoBites
ExpertActivity

Activity

The Activity tab gives you real-time operational metrics for Expert, query volume, hit rate, latency, and active integrations. Use it to verify your integrations are working and monitor health.


Overview

Activity shows everything happening with your Expert in real time. Every query from every connected agent is logged with its source, result count, and latency. This is the first place to check when debugging an integration or monitoring adoption.


Metric cards

Four cards at the top of the Activity tab summarize your Expert’s health:

Query volume

Total number of queries received in the selected time period. Shows how actively your agents are using Expert.

Hit rate

Percentage of queries that returned at least one result. A low hit rate means you may need more content, check the Intelligence tab for gaps.

Latency

Average response time in milliseconds. Expert typically responds in under 500ms. Spikes may indicate infrastructure issues.

Active integrations

Number of distinct integrations that sent queries in the selected period. Helps you confirm all connected agents are active.


Volume chart

Below the metric cards, a time-series chart shows query volume over time. You can filter by:

  • Integration, see queries from a specific agent (Intercom, Zendesk, Claude, etc.).
  • Time period, last 24 hours, 7 days, 30 days, or custom range.

Use the chart to spot trends, increasing volume means growing adoption, sudden drops may indicate a broken integration.


Recent queries

A real-time log of individual queries with details for each:

  • The exact query text sent by the agent.
  • Which integration sent it.
  • Number of results returned.
  • Response latency.
  • Timestamp.

This log is essential when setting up a new integration, you can immediately confirm that queries are arriving and returning results.


What to look for

Verify integrations are connected

After setting up a new integration, check Activity to confirm queries are arriving. If not, double-check the API key and endpoint configuration.

Spot issues early

A sudden drop in volume or spike in latency can indicate a configuration change or infrastructure issue. Activity surfaces these before your users notice.

Monitor hit rate

A declining hit rate means more queries are returning empty results. Head to the Intelligence tab to see which queries are failing and create content to fill the gaps.