AI Insights
AI-generated summaries of review sentiment themes and GBP performance patterns across the client portfolio.
What It Shows
AI Insights generates natural-language summaries of GBP performance patterns and review sentiment themes, surfacing actionable observations that would take too long to derive manually from raw data.
Insight types generated:
- Sentiment theme summary — common themes in positive and negative reviews (e.g. "Staff friendliness mentioned in 68% of 5-star reviews")
- Performance anomaly — flags unusual drops or spikes in impressions, calls, or clicks
- Competitive gap — identifies locations with declining review velocity compared to portfolio peers
- Review response gap — highlights locations with low response rates
How Insights Are Generated
Insights are generated weekly. The pipeline aggregates review text and performance metrics per client, then prompts the LLM to identify patterns and generate structured observations.
Each insight includes:
- A headline observation (1 sentence)
- Supporting data (metric, trend direction, comparison period)
- A suggested action for the account manager
Insight Confidence
Not every auto-generated insight is actionable. The AI Insights view allows account managers to mark insights as "Helpful" or "Not relevant." This feedback loop is stored and will be used to tune the insight generation prompts over time.