Analytica

Ask your data. Understand your business.

Governed AI analytics for teams that need trustworthy answers.

Natural-language questions become validated QueryPlans — never arbitrary SQL from the model — so every insight stays evidence-backed and tenant-safe. The live demo is a working sample prototype, not a production tenant.

Prototype uses sample commerce tenants and works end-to-end as specified.

How it works

From question to governed insight in one loop.

The same hero journey proven in the SDD Analytics Copilot prototype — productised for a multi-tenant SaaS.

  1. 01

    Ask

    Pose a business question in plain language — revenue, regions, segments, trends.

  2. 02

    Plan

    The agent proposes a QueryPlan against your semantic metrics and dimensions.

  3. 03

    Execute

    The semantic engine validates, compiles, and runs parameterised queries with tenant scope injected.

  4. 04

    Explain

    Answers arrive with evidence, charts, and a trace you can inspect — then pin to a dashboard.

Governance first

AI decides what to try. The semantic engine decides what is legal.

Analytica never lets the model execute raw SQL. Plans are validated against your metrics layer, compiled to parameterised queries, and always scoped to the authenticated tenant.

{
  "metric": "revenue",
  "dimensions": ["region"],
  "time": {
    "dimension": "order_date",
    "range": "last_month"
  },
  "comparison": "previous_period"
}

QueryPlan contract — the hard boundary between reasoning and execution.

Implementation

Talk to us about Analytica

Interested in a governed analytics Copilot for your stack? Send a note — you’ll get an acknowledgement email, and we’ll follow up personally.