Sample brief · illustrative data

A week in the life of an agent fleet.

This demo fleet's weekly brief shows fleet totals and 6 selected agents, ranked by attention. The numbers are illustrative; the behavior is the product's.

fare-reconciliation-agent needs a look first: labeled success 45.2% (-8.1 pts).

Spend: seat-booking-agent is over its monthly budget and waiting for your sign-off. Spent $47.06 of a $37.65 plan for the period that started Oct 1.

This week · 7 days to 2026-10-25

Demo fleet totals

These totals cover the full demo fleet, including additional demo agents not shown in the selected rows below.

Spend

$538.53

6% up on the previous week ($507.37).

Steady on last week

Agent runs

3,988

1% more than the previous week (3,939).

Across 7 of 7 days

Production success rate

79.2%

2,857 runs have a known outcome (70% of runs). -1.6 points on the previous week (80.8%).

Worth watching

Selected agents, most urgent first

The 6 selected agents below are a subset of the fleet. Their runs, labeled outcomes, and spend do not add up to the fleet totals above. Each success rate uses only that agent's labeled runs.

  1. fare-reconciliation-agentNeeds attention
    Labeled runs
    610 of 773
    Success
    45.2%
    Vs last week
    -8.1 pts
    Spend
    $68.80
  2. triage-agentNeeds attention
    Labeled runs
    397 of 397
    Success
    78.8%
    Vs last week
    -11.8 pts
    Spend
    $90.54
  3. seat-booking-agentNeeds attention
    Labeled runs
    85 of 85
    Success
    96.5%
    Vs last week
    +7.4 pts
    Spend
    $12.00

    From latency threshold alert firing, not labeled success

  4. fare-rules-agentWorth watching
    Labeled runs
    332 of 438
    Success
    72.9%
    Vs last week
    +0.0 pts
    Spend
    $61.87
  5. baggage-claims-agentNot enough labeled outcomes
    Labeled runs
    0 of 700
    Success
    —
    Vs last week
    —
    Spend
    $46.90
  6. loyalty-agentHealthy
    Labeled runs
    1,161 of 1,400
    Success
    94.2%
    Vs last week
    +1.8 pts
    Spend
    $59.74

Verdicts: Needs attention Worth watching Not enough labeled outcomes Healthy

An alert fired this week

Investigation · Latency threshold

seat-booking-agent

Needs review

Couldn't form a confident hypothesis — here's the raw correlated data.

  • metricp95 latency 61.8s over the last 20 runs before the alert, against the 8s threshold; 53.6s over the 7 days before.
  • traceThe slowest run before the alert.

Correlated data is attached; a provider-side latency incident would explain the shape but could not be confirmed from your traces.

Confidence: Low · Investigation cost $0.0041

Read how the brief is built →Join the waitlist