Detect the anomaly
ACAI monitors every KPI continuously. When activation, retention, or conversion shifts beyond the baseline, it triggers an investigation automatically.
AI Root-Cause Analysis
Find why metrics changedwithout hours of digging.
When a KPI drops or spikes, ACAI automatically isolates the affected segment, traces the causal event sequence, and delivers a recommended fix — grounded in your real clickstream data.
AI Root-Cause AnalysisTriggered automatically
⚠ Churn spike detected — +8.3%
March cohort vs February baseline — investigation started
01
Anomaly detected
Activation dropped 8.3% — March cohort vs Feb baseline
02
Segment isolated
Users who skipped onboarding step 3 — 1,284 accounts
03
Cause identified
72% exit before completing the setup checklist
04
Fix recommended
Add guided tooltip — est. impact: −40% churn next cohort
Mean time to root cause
< 2 minvs days of manual investigation
Causal layers examined
7+segment, event, device, time, source, cohort, flow
Fix accuracy
91%of AI-recommended fixes confirmed by teams post-ship
How it works
From your first event to your first answer.
Isolate the affected segment
The analysis engine narrows down which users, devices, cohorts, or traffic sources are responsible for the change — and which are unaffected.
Trace the causal event chain
ACAI replays the event sequences of affected versus unaffected users to identify the exact step, screen, or interaction that diverges.
Deliver the fix recommendation
You receive a written root-cause summary, the supporting evidence, and a concrete recommended action with estimated impact.
Capabilities
Everything the feature does.
Automated anomaly triggers
No manual setup. ACAI watches every tracked KPI and begins investigation when movement exceeds configurable thresholds.
Multi-dimensional isolation
Segment isolation across device, OS, traffic source, plan tier, geography, and any custom user property simultaneously.
Event sequence replay
Compare the exact clickstream path of affected versus control users to find where the journey diverges.
Counterfactual analysis
ACAI models what the metric would have looked like without the causal factor to quantify the true impact.
Fix impact estimation
Before you ship a fix, ACAI estimates the expected metric recovery based on historical patterns and affected user volume.
Audit trail
Every investigation is logged with full evidence so you can review, share, or revisit the analysis at any time.
Root-Cause Intelligence
From metric drop to fix recommendation in under two minutes.
Manual investigation means days of pulling cohorts, replaying sessions, and guessing. ACAI automates the entire causal chain — isolation, evidence, and a recommended action with estimated impact.
Auto-triggeredCausal evidenceFix estimationFull audit trail
Root causes found
94%of anomalies
ACAI successfully identifies a root cause for 94% of detected metric anomalies.
1 / 3
Use Cases
When teams reach for this feature.
Activation dropped 8% — what happened?
ACAI traces the drop to users who skipped onboarding step 3, shows the exact event divergence, and recommends a tooltip fix.
D7 retention fell in the March cohort
ACAI isolates the cohort, finds the missing habit-forming action, and estimates a 40% retention recovery with one product change.
Checkout conversion dipped on mobile
ACAI compares mobile versus desktop event sequences, identifies the payment-step friction, and quantifies the revenue impact.
Resolve metric drops in minutes, not days
Eliminate guesswork with evidence-grounded causal chains
Get a recommended fix with estimated impact before you ship
Build a searchable investigation history for future reference