Ask your business data in plain English.

Not another SQL-writing chatbot

Most “AI analyst” tools let a model write SQL and hope it’s right. That’s unreproducible and impossible to audit. Sluice draws a hard line: the AI plans — it fills a structured KPI (measure, dimensions, filters, joins) over your real schema — and a deterministic compiler writes every query. The same question always compiles to the same SQL, and every number is checkable against the raw data.

Describe your business. Get a plan.

You don’t start by modeling data. You start by describing what your business does. The analyst reads your connected sources and proposes:

• Entities & relationships — what each table represents and how they join.
• KPIs — the measures your business actually cares about.
• A dashboard — headline numbers, breakdowns and trends, ready to approve.

You review and approve the plan. Then you just ask.

How a question becomes an answer

Answers you can act on

Because the analyst understands your relationships, it resolves business language — “customer”, “active”, “overdue”, “margin” — to the right tables and builds fan-out-safe joins across them. Multi-metric questions produce a rich, multi-chart answer, not a single flat number. And every figure remains traceable to source, so you can put it in front of a customer, an auditor, or your board.