Ten specialized analyzers that proactively flag spend anomalies, deadline risks, and budget issues before they become problems.
The detection layer runs ten independent analyzers across financial domains: profitability (margin compression and net loss), revenue (period-over-period and year-over-year decline), expenses (category concentration, vendor outliers, statistical spikes), cashflow (overdraft, low reserves, burn rate, projected shortfall), receivables (overdue concentration, slow payers, client concentration), budget (overspend and category-level overruns), ad-spend daily-spend spikes (per-client/platform), ad-spend pacing and delivery health (underspend, overspend, stopped, paused-with-budget, stale-sync, zero-conversion), per-client (scope creep and revenue concentration), and transaction-level outliers. Each analyzer runs in parallel during scheduled and on-demand scans.
Anomalies are surfaced proactively through the Activity Hub and AI chat without requiring anyone to run a report or ask a question. When the spend analyzer detects that a campaign burned through 60% of its monthly budget in the first week, a notification appears immediately with the relevant context and suggested actions. This turns your team from reactive (discovering problems at month-end) to proactive.
Each anomaly is scored as critical, warning, or info based on rule-based thresholds matched to financial impact and time sensitivity. Critical anomalies (e.g. a net-loss period or a projected cash shortfall) trigger Smart Watch + email notifications immediately. Warning and info findings appear on the Anomalies page and in the daily digest, but do not page anyone overnight.
Each detected anomaly becomes a persistent incident with status (open / acknowledged / snoozed / resolved / dismissed) and a full audit trail. Snoozing buys time without losing the signal — when the snooze expires, the row flips back to open if the underlying issue is still detected. Resolved-and-recurring anomalies create new incidents rather than reopening old ones, preserving incident history. Correlated findings (for example a low-margin month with margin compression and revenue decline) collapse under a single parent incident card.