Console

Welcome back, Alex

Your AI platform is healthy. Traffic is up and spend is tracking under budget.

View report

Requests (24h)

128.4k

↑ 12.3% vs last week

Tokens used

84.2M

↑ 8.1% vs last week

Spend (MTD)

$6,240

↓ 3.4% vs last week

Avg latency

412 ms

↑ 18 ms vs last week

AI insights

Every insight here is sample data shipped with the component. None of it is about your system.

AI-Generated Insights

6 new

Critical

94%

confidence

Churn Risk

Churn Risk: 12 Enterprise Accounts

Usage patterns indicate elevated churn probability. These accounts show a 40%+ drop in feature engagement over the past 14 days, correlating with historical pre-churn signals.

$284KAt-risk ARR−43%Engagement drop

8m ago · Churn Predictor v3.2

94% — no method or model recordedNo evidence linked — this is an assertion, not a finding.

High

87%

confidence

Revenue

Cross-Sell Opportunity: 234 Accounts

Growth tier users with high API usage and no add-ons have a 3.4× higher upgrade propensity score. Historical data shows this cohort converts at 28% when contacted within 7 days.

+$41KPotential MRR3.4×Avg propensity

34m ago · Revenue Intelligence v2.0

87% — no method or model recordedNo evidence linked — this is an assertion, not a finding.

High

91%

confidence

Performance

API Latency Anomaly — EMEA

p95 response time in the EMEA region spiked to 680ms, 4.8× above the baseline of 142ms. The anomaly started at 09:14 UTC and correlates with a recent worker queue deployment.

680msp95 latency142msBaseline

52m ago · Anomaly Detection v4.1

91% — no method or model recordedNo evidence linked — this is an assertion, not a finding.

Medium

78%

confidence

Budget

Q3 Marketing Budget Overrun Forecast

Current Q3 marketing spend trajectory projects a $64K overrun (+18%) against the approved $360K budget. Paid social and content channels are the primary contributors.

+$64KProjected overrun$298KSpend to date

2h ago · Budget Forecaster v1.4

78% — no method or model recordedNo evidence linked — this is an assertion, not a finding.

Medium

82%

confidence

Engagement

Export Feature: Hidden Growth Signal

New Pro users who discover the bulk-export feature within the first 7 days have a 3× higher 90-day retention rate. Only 12% of new Pro users currently find it without guidance.

+3×Retention lift12%Current discovery

5h ago · Product Analytics v2.3

82% — no method or model recordedNo evidence linked — this is an assertion, not a finding.

Low

69%

confidence

Security

Unusual Admin Login Patterns

3 admin accounts authenticated from new geographic regions outside their established baselines. All events occurred within a 40-minute window — possible credential compromise.

3Affected accounts4New geos

8h ago · Security Monitor v1.8

69% — no method or model recordedNo evidence linked — this is an assertion, not a finding.

Token usage

Open usage

Token Usage Overview

Last 30 days · All models

Figures are list prices. They exclude any contract discount and tax, so the invoice will be higher or lower. Recent usage may not be settled by the provider yet.

Live tracking

+1.1%

3.06M

Total tokens

+1.3%

$52.34

Total cost

vs period

9,008

API requests

per req

$0.0058

Avg cost / request

64th pct

Token Distribution

3.06M total

Input

1.98M

65% of total

Output

1.07M

35% of total

19% of these models' input was served from cache and billed at a lower rate.

8% of these models' billed output is reasoning nobody sees.

Output tokens cost ~3× more than input tokens

By Model

$47.42 total

GPT-4o

OpenAI

1.20M $22.40

43%

GPT-3.5 Turbo

OpenAI

722.8K $0.72

26%

Claude 3.5 Sonnet

Anthropic

582.8K $9.48

21%

Gemini 1.5 Pro

Google

278.5K $14.82

10%