Product · Context Graph
Give the next question more context than the first.
The Context Graph combines semantic structure with the business concepts, decisions, clarifications, and analytical patterns needed to interpret a real question. It keeps that meaning outside a one-off prompt so people can inspect what informed the plan.
Much more than a semantic layer
A semantic layer describes the data.
Company context explains the question.
The Context Graph is broader than a metric catalog or semantic layer. It brings together company language, analytical choices, model structure, and knowledge supplied by people during the work—including meaning that was never modeled in advance.
Business concepts
Terms people use to ask the question.
Segments, measures, events, cohorts, thresholds, and other company-specific language can be represented explicitly.
Analytical structure
How a recurring analysis is shaped.
Calculations, comparison patterns, scope, grain, and related analytical choices can inform a later plan.
Data relationships
What is present in the configured model.
Tables, columns, keys, joins, measures, and supported metadata sources ground interpretation in the connected data estate.
Interaction context
What the team supplied during analysis.
A definition or clarification supplied in a workflow can remain available as workspace context when that behavior is enabled and tested.
Context in the plan
Origin stays visible beside the meaning.
A plan should not flatten what the team defined, what the model contains, and what Spotonix assembled into one anonymous block.
Which product categories do priority customers prefer?
Interpretation
priority customers
Customers with at least five orders across store, web, and catalog channels.
Reused from workspace contextAnalytical plan
- Segment
- Priority customersReused from workspace context
- Group by
- Product categoryFound in your model
- Measure
- Purchase countAssembled by Spotonix
- Grain
- Customer × categoryAssembled by Spotonix
Read the interpretation and plan. Missing meaning is visible before the analysis runs.
Inspect the basis of the result. The answer does not have to hide behind a chat response.
What it does not imply
Explicit context is useful because its boundary is explicit too.
The Context Graph is not a claim that Spotonix owns the definition, stores every data row, or resolves every organizational conflict.
Not a data copy
The graph does not replace the warehouse.
It represents context used to interpret analysis. Retained artifacts, storage location, and network boundaries are documented for the deployment.
Not model weights
Company meaning remains inspectable structure.
A person can read and correct an explicit concept without retraining the underlying language model.
Not automatic authority
A supplied term does not appoint its own owner.
Origin labels identify where context came from. Author-level provenance, conflict policy, and organization-wide authorization must be evaluated separately.
Not guaranteed memory
Persistence and retrieval are workflow behaviors.
Test whether the right context remains available, how updates are handled, and when a later question asks again.
Continue the technical path
See how context becomes a plan—and how the system is bounded.
The interpretation mechanism
Intent Algebra
How Spotonix represents selected meaning, scope, filters, grain, and ambiguity before execution.
The system boundary
Architecture
What is stable in the product path and what must be confirmed for the deployment.
The input boundary
Integrations
Which warehouse, model, and metadata paths are evidenced today and which remain engagement-specific.
Evaluate directly
Test whether the right meaning carries forward.
Include repeated questions, changed definitions, conflicting terms, and reviewer handoffs in the proof—not only the happy path.
Book DemoSee how it works