Builds from validated Segments and Calculations.
Every answer is composed from your team's already-accepted business logic — never generated from scratch.
Composes the Plan
Builds answers from validated logic. Not from scratch.
Spotonix composes every answer from your team's already-accepted Segments and Calculations. The plan is visible before any SQL runs.
What Composes Means
The plan is the trust mechanism.
LLM-only stacks generate SQL token-by-token, hoping it's right. Spotonix proposes a symbolic plan in BI Algebra first — composed over validated building blocks — and only compiles SQL after the plan is accepted. The plan is the contract between human intent and machine execution.
The plan is visible before any SQL runs.
You see exactly which Segments, Calculations, and dimensions Spotonix chose — and what each one resolved to.
Algebra closes, or the system asks.
If a concept is ambiguous, Spotonix surfaces the choice. No silent guess. No probabilistic SQL.
Same plan compiles to the same SQL. Forever.
Deterministic compilation means the same question Monday and Tuesday produces byte-identical SQL — and therefore the same number.
Composition Workflow
From proposed plan to deterministic answer.
The LLM is useful, but it is not the source of truth. The accepted plan is what the system verifies, reuses, and compiles.
Propose the plan.
The LLM proposes a symbolic plan in BI Algebra, composed over your validated Segments and Calculations — not free-form SQL.
Close the algebra.
The system accepts the plan only if every concept resolves. Ambiguity is surfaced for human-in-the-loop clarification.
Compile and explain.
Once accepted, SQL compiles deterministically. The plan stays attached to the answer as its audit trail.
context.compose
Plan Habitual Buyers + QoQ Customer Count Change + Store × Quarter
Each line points to a named, validated building block in your Context Graph.
Verification Algebra closes ✓
Every concept resolves to a definition your team accepted.
Execution Same plan → same SQL → same answer
Deterministic compilation. No drift between sessions, analysts, or LLM versions.
Why It Matters
Composing is the trust mechanism.
Data teams can defend the number.
Every answer carries its plan. Every concept points back to the validation that accepted it.
Compliance has something to audit.
Definitions, assumptions, and lineage attach to the answer — not buried in a chat log.
Business users understand the why.
The answer isn't a number in a chart. It's a number with a readable explanation of how it was computed.
The same question stays stable.
The accepted plan is reusable. Next week's identical question doesn't improvise a different query.