Explaining Changes via AI
The AI assistant analyzes changes in metrics and suggests possible causes — from technical failures to changes in campaigns.
How diagnostics work
- The AI retrieves data for the current and previous periods
- It calculates the delta and determines whether the change is anomalous
- It analyzes possible factors: campaign changes, seasonality, technical issues
- It provides an explanation with potential causes and recommendations
Typical diagnostic questions
Metric drops
- "Why did conversions drop this week?"
- "Why did CTR decline in Meta Ads over the last 3 days?"
- "What happened to spend in Google Ads — did it drop sharply?"
Metric growth
- "Why did cost per lead increase in Google Ads?"
- "What caused the sharp rise in impressions on TikTok?"
- "Why did CPA increase by 40% over the last week?"
Finding the worst/best performers
- "Which campaign has the worst CTR over the last 7 days?"
- "Which source has the highest cost per click?"
- "Where is the lowest ROAS across all channels?"
What the AI analyzes
| Factor | What's checked |
|---|---|
| Budget changes | A sharp increase/decrease in a campaign's daily budget |
| Seasonality | Comparison with the same period last year |
| Competition | Rising CPM/CPC as a sign of increased auction pressure |
| Technical failures | API errors, tracking outages, pixel issues |
| Creative changes | Ad creatives going stale or being updated |
| Campaign status | Paused, ended, or rejected by moderation |
Diagnostic limitations
- The AI only sees data available through the API — it doesn't analyze creatives visually
- Diagnostics work best on periods of 3 days or more
- Accurate diagnosis of attribution changes requires a configured attribution model
- The AI suggests probable causes, not guaranteed ones — the final call is yours