Asking Questions All Users
The simplest way to use AI Chat is to ask it a question in plain language. No special syntax or query format is needed.
Example Questions
| What you ask | What you get |
|---|---|
| ”Show me revenue this month” | A metrics card with total revenue and a trend chart. |
| ”Compare last 7 days vs previous” | Side-by-side KPIs with delta percentages. |
| ”Which campaigns have the highest eCPM?” | A ranked table of campaigns sorted by eCPM. |
| ”What’s unusual in yesterday’s data?” | An anomaly summary highlighting metrics that deviate from recent trends. |
| ”Break down impressions by country” | A geographic breakdown table or chart. |
| ”How is campaign X performing?” | A focused performance summary for that specific campaign. |
Response Formats
AI Chat does not just return text. Depending on the question, responses can include:
- Metrics cards — Single-value KPIs with period comparisons.
- Tables — Sortable rows with relevant columns.
- Charts — Line, bar, or area charts for trends and distributions.
- Text summaries — Plain-language explanations of what the data shows.
These elements appear inline in the chat. You can scroll through them, and they update if you change the period selector.
Tips for Better Answers
- Be specific about time. “Revenue this month” is clearer than “revenue.”
- Name the metric. “eCPM” or “fill rate” gets a direct answer. “How are things going” gets a general summary.
- Reference campaigns by name. The assistant can look up specific campaigns if you mention them.
- Ask follow-ups. The chat remembers context within a conversation, so you can say “now show that by day” without repeating the full question.
The assistant respects your role permissions. It will only show data you have access to in your current workspace.