AI Insights

Pishoniq goes beyond reporting numbers. It interprets your event stream and surfaces findings in language your team can act on.

How it works

Pishoniq continuously analyzes your events for:

• Drop-offs — Declines in conversions, signups, or retention
• Anomalies — Unusual changes compared to your established baseline
• Funnel breaks — Steps where users abandon at higher-than-expected rates
• Cohort shifts — Changes in behavior across user segments

When a meaningful pattern is detected, an insight appears on your dashboard with a clear explanation.

Natural language queries

Open the AI Chat in the dashboard and ask a question:

• "Why did signups decline last week?"
• "Which acquisition channel delivers the highest retention?"
• "Which features do churned users rarely engage with?"

Pishoniq returns a written analysis supported by your data, with a chart and the specific user cohort that drove the change.

Suggested insights

On the dashboard home, Pishoniq shows a list of suggested questions tailored to your data — based on what's actually interesting in your event stream right now. Click any suggestion to expand the analysis.

Anomaly detection

Pishoniq automatically detects:

| Anomaly | What it means | |---|---| | Conversion drop | A funnel step lost more than expected conversions vs. baseline | | Traffic spike | A page received significantly more sessions than usual | | Retention cliff | A cohort is churning earlier than historical cohorts | | Feature abandonment | A previously popular event has gone quiet |

Each anomaly shows the affected metric, the magnitude of the change, and the cohort that drove it.

Funnel analysis

Track a conversion sequence:

PishonIQ.track('signup_started')
PishonIQ.track('email_verified')
PishonIQ.track('onboarding_completed')
PishonIQ.track('subscription_created')

Open the Funnels page in the dashboard, define the steps, and Pishoniq maps the drop-off between each step — including commentary on the biggest leak.

Churn signals

Pishoniq monitors behavioral signals associated with disengagement:

• Reduced session frequency
• Feature abandonment (events that used to fire no longer do)
• Drop in core action events (e.g., fewer feature_used events per session)

When these signals cross a threshold, an alert fires so your team can intervene while there's still time.

Recommendations for better insights

1. Track consistently — A stable event schema improves analysis quality over time.
2. Identify users — Calling PishonIQ.identify() unlocks retention, cohorts, and per-user analysis.
3. Enrich with traits — Plan, role, and acquisition source enable more precise cohort explanations.
4. Review insights weekly — Many teams spend a few minutes each week scanning new findings.

Support

For assistance, reach out via the in-app chat or email support@pishoniq.com.