Skip to main content

​
Overview

Insights lets you chart events, user profiles, and cohorts over time. Use it for adoption, activation, engagement, and segment comparisons in AI and SaaS products.

​
Use cases

  • How is weekly active usage trending?
  • How many agent_run_completed events per day?
  • Which templates or workflows drive the most runs?
  • Which campaigns drive signups and activations?
  • How does paid compare to organic for qualified trials?
  • Upgrades and revenue events (trackCharge) over time
  • Plan mix by segment

​
Quick start

1

Define your metric

Add metric and choose:
  • Event — counts or aggregates from track() / auto-events
  • Profile — user property rollups
  • Cohort — cohort size or overlap (where supported)
2

Choose measurement

  • Unique users
  • Total count of events
  • Property sum / average on events or profiles
  • Sessions (where available)
3

Apply filters

Event properties, user properties, time range
4

Add breakdowns

e.g. plan, country, workflow_id, surface
5

Choose visualization

Line for trends, bar for comparison, table for detail

​
Measurements (common)

​
Event-based

MeasurementDescriptionExample
Unique usersDistinct users who fired the eventWho completed a run
Total countEvent occurrencesTotal runs
Event property sum / avgNumeric property on the eventSum of duration_seconds

​
Profile-based

MeasurementDescriptionExample
Profile countUsers matching criteriaAccounts on enterprise plan
Property sum / avgRollup on people propertiesSeats, credits, LTV fields you store

​
Filters

​
Event filters

event_name = "agent_run_completed"
AND outcome = "success"
AND last 30 days

​
User property filters

plan = "pro"
AND country = "US"
See Filters & operators.

​
Breakdowns

Segment a metric by a property: 
agent_run_completed
breakdown: workflow_id

→ Compare volume per workflow
You can use event properties, user properties, and time buckets (day, week, month).

​
Visualizations

ChartBest for
LineTrends
Column / barComparisons and rankings
StackedPart-of-whole over categories
PieSimple share (use sparingly)
TableMany dimensions

​
Advanced

​
Multiple metrics

Plot signup_completed, agent_run_completed, and subscription_upgraded together to relate top-of-funnel to depth and revenue.

​
Formulas

Build ratios such as activation rate: users who fired activation event ÷ users who signed up (using two metrics and a formula, where the product supports it).

​
Time comparison

Compare this week to last week on the same metric for quick regressions.

​
Examples

​
Runs per day

  • Event: agent_run_completed
  • Measurement: Total count
  • Time: Last 30 days, daily
  • Chart: Line

​
Activation by campaign

  • Event: activation_milestone_hit
  • Measurement: Unique users
  • Breakdown: utm_source
  • Chart: Bar

​
Time in workflow

  • Event: agent_run_completed
  • Measurement: Average of duration_seconds
  • Breakdown: workflow_id

​
Best practices

  • Start with one metric and a clear time range, then add breakdowns.
  • Name saved reports by decision: e.g. “Weekly runs by workflow”.
  • Align event names with naming conventions.

​
Saving and sharing

Save reports, add them to Boards, and share with your team per workspace permissions.

​
Next steps

Funnels

Step-by-step conversion

Filters reference

Operators and examples