Chat & AI agents
Conversational analytics — ask the AI a question and get an answer grounded in your data. Build charts, run playbooks, surface anomalies.
Last updated
Conversational analytics — ask the AI a question and get an answer grounded in your data. Build charts, run playbooks, surface anomalies.
SMAQ AI isn't a generic chatbot stitched to your data. It's a team of specialized agents — fetch-data, visualization, search-datablock, business analysis, report writing — coordinated by a supervisor that routes your question to the right one.
[Screenshot needed] Chat panel with a multi-turn conversation showing a question, the AI's thinking task list, a generated chart, and follow-up turns.
Project chat — click Chat in the side menu. The AI has access to every datablock, dashboard, automation, and KB entry in the project.
Dashboard chat — the chat icon in any dashboard header. Context is scoped to the current dashboard.
Public dashboard chat — anonymous viewers on a public dashboard can chat about the data they see.
What SMAQ AI can do — full agent reference, capability list, examples.
Playbooks — codify reusable analyses the AI can run on demand.
Be specific. "ROAS by campaign last week" beats "campaign performance."
Name the source. "Google Ads spend" beats "spend" if you have multiple ad platforms.
Use a date window. "Last 30 days" beats "lately."
Iterate. Treat it like a conversation. "Now break that down by device." "Filter to just search campaigns."
Add to dashboard — when the AI returns a chart you like, push it onto a dashboard so it's persistent.
Chat conversations are private to your project by default.
Organization Owners can view all historical messages — see Review historical user messages.
SMAQ's AI runs on Anthropic's Claude. See SMAQ's privacy policy for processing details.
Every chat message consumes credits. Complex multi-step queries (multiple data fetches, AI analyses) cost more than single-fact questions. See Credits & top-ups.
Last updated