The full flow
- Data pull: scheduled daily (also manually triggerable) from your connected Google / Meta / TikTok accounts — 7-day and 30-day metric snapshots
- Anomaly detection: against historical baselines, flag CPA > target by 30%, ROAS < 1.5, CTR < 50% of industry average, etc.
- LLM analysis: anomalies + account metadata + prior optimization history → Claude (Sonnet 4.6) or OpenAI GPT-4o
- Structured output: the model writes 3–8 specific recommendations via our JSON schema — each with type / target / reason / expected impact / confidence
- You decide → you make the change: recommendations appear in your dashboard as advice only; you decide whether and how to act, and make the change yourself in Google Ads / Meta Ads Manager / TikTok Ads Manager. The AI never changes your account on its own
Recommendation types
- pause_keyword — pause high-CPA keywords
- pause_campaign — pause an entire losing campaign
- increase_budget — push more budget to high-ROAS ad groups
- decrease_budget — cut budget on inefficient ads
- add_keyword — add keywords surfaced from search-term reports
- change_bid — adjust bid strategy or target CPA / ROAS
- creative_refresh — creative fatigue → swap assets
Three confidence tiers
- high: clear data outliers (e.g. CPA $45 vs target $8) — consider these first
- medium: directionally right but uncertain (e.g. seasonal effects) — small-step recommended
- low: advisory only — human review recommended
Recommendations never take effect on their own — your account changes only when you make the change in your ad platform. If results disappoint, change the setting back in your ad platform; AdWhiz itself has no rollback feature.