// you
Which customers are at risk this month?
// agent
42 customers flagged · usage down >40% week-over-week · top 5 below.
Ask anything about your business, in plain English. The Analyst Agent runs on the same Domain Model as your pipelines, dashboards and reports — the AI's map of your business: every metric defined once, every entity and join, every access rule. It returns charts and data snippets inside the conversation, and pins any of them to a dashboard with a single click. Runs inside your infrastructure, on the LLM you approve — nothing leaves your boundary. Embeddable in your own product, with your branding.
The Analyst Agent isn't a search bar. It reasons about your business the way a senior analyst would — picks the right metric, the right cut, the right chart, and shows its working alongside the answer.
"What's driving APAC churn this month?" "Compare Q3 to Q2 for Enterprise customers." The agent figures out the join, the grain and the time window.
Every answer ships with a generated chart or data table — right in the conversation. No "let me build a dashboard" detour, no spreadsheet export.
Liked the chart? Pin it to a new or existing dashboard — it lands governed, scheduled and ready to share. The exploration becomes the artifact.
The first question every team asks about AI analytics is how they would know if it were wrong. Sprinkle answers it four ways, on every answer. And when the data does not support an answer, the agent says so instead of filling the gap.
Metrics come from the Domain Model. If the agent writes its own aggregate over a governed metric, the query is rejected before it runs, so the answer uses the same formula as the dashboard.
Every sum, count and ranking runs as SQL against your data. The agent reads the result. It does not do the maths from a preview of rows.
Open the query, see which tables and formulas it used, and download the rows behind the chart. Verify rather than trust.
Answer quality is scored against reference questions whose answers are computed by SQL, so a change of prompt or model shows up as a score, not a surprise.
The Analyst Agent reads the same Domain Model your dashboards and pipelines do — entities, joins, governed metrics, glossary terms. When someone asks about "MRR," the agent uses the one definition your finance team signed off on. Same numbers as the Finance dashboard, every time.
The Analyst Agent calls the model your organisation has approved — hosted in your cloud account or self-hosted on your infrastructure. Prompts, schema context and results stay inside your boundary, and nothing goes to a model vendor you have not chosen.
Most ad-hoc analysis dies in a Slack thread. Here, every chart the agent generates is one click away from being a governed widget on a real dashboard — same grounding, same refresh, ready to share with the team.
Give your customers the same Analyst Agent — themed to your brand, scoped to their tenant. They ask questions in plain English, see charts in chat, and pin insights to dashboards that live in your product, not ours.