Ask your data anything
A conversational analyst for your governed data. Ask in plain English and get charts, narratives, and full dashboards back — with the exact query visible and every permission enforced, so you can trust every answer.
- Plain-language questions become charts, dashboards, and reports
- Explainable: see the exact SQL behind every answer
- Governed: runs with your permissions, answers from certified metrics
Which region grew fastest this quarter?
APAC grew +31%, driven by enterprise expansion:
▸ view generated SQL
Watch it think
From question to built dashboard
Ask in plain language; the analyst plans, runs governed queries, and builds the chart — showing its work at every step.
Reasoning
- Resolve the certified “Revenue vs Target” metric
- Check row-level security for you
- Evaluate by region, flag anything under 100%
A closer look
Inside AI Analyst
Ask, and it builds
More than answers: a governed agent with 40+ actions goes from a plain-language question to saved charts, dashboards, and reports — showing a plan you approve before anything changes.
- 40+ governed tools: query, chart, dashboard, report, alert
- A plan preview and confirm before it writes anything
- Evaluates your certified metrics, not re-derived numbers
Governed, and explainable
The analyst authenticates as you and answers within your row-level security — failing closed when access is uncertain. Every answer can show the exact SQL it ran.
- Runs with the asking user's permissions and RLS
- Exact SQL surfaced and logged for every answer
- Anomaly, key-influencer, and forecast insights, all inspectable
Insight without a key
Scheduled digests deliver keyless, statistical on-pace and behind-target projections for your KPIs — and a built-in MCP server opens your governed data to your own AI tools, read-only.
- Keyless pace-to-threshold digests to your channels
- Glossary-aware answers in your business language
- MCP server: your AI stack queries governed data safely
What you can do with AI Analyst
Ask, and it builds
More than answers: a governed agent with 40+ actions goes from a plain-language question to saved charts, dashboards, and reports — showing a plan you approve before anything changes.
Explainable answers
Every response can show the exact SQL it ran and the metric definitions it used. Numbers come from real queries against your data, never the model's memory.
Governed grounding
The analyst authenticates as you and answers within your row-level security — failing closed when access is uncertain — using your certified metrics so figures match your dashboards.
Statistical insight
Built-in anomaly detection, key-influencer analysis, forecasts, and 'explain this change' delta decomposition — all deterministic and inspectable, not guesses.
Automated digests
Scheduled metric digests deliver keyless on-pace and behind-target projections for your KPIs to your team's channels — insight that ships with no AI key required.
Open to your AI stack
A built-in Model Context Protocol (MCP) server lets your own AI tools query your governed data read-only, with the same permissions and row-level security enforced.
40+ actions
A governed agent that acts
Explainable
SQL shown for every answer
MCP-ready
Open to your own AI tools
How it works
Ask a question
"Why did revenue drop last week?"
Review the answer
Chart, narrative, and the query behind it
Dig deeper
Ask follow-ups; context carries forward
Save to a dashboard
Pin any answer as a live card
Frequently asked questions
A frontier large language model behind a provider-neutral gateway, orchestrated with strict guardrails: the analyst can only act through governed tools, with the asking user's permissions and row-level security.