dbt
in the intelligence stack.

What dbt documents

dbt provides tools for building, testing, documenting, and collaborating on data transformations. Its analytics-engineering approach applies software-development practices to the models that make warehouse data usable.

This profile was reviewed against the official sources listed below on September 26, 2026. Product capabilities, names, licensing, availability, and policies can change; verify current details with the provider before making a purchasing or architecture decision.

Where the platform is strong

• Version-controlled transformation logic close to analytical data
• Testing, documentation, dependency, lineage, and reusable modeling practices
• Strong fit for transparent business models and collaborative analytics engineering
• Broad warehouse and data-platform ecosystem

A strength is not a universal recommendation. It describes where the documented product model can create leverage when the implementation, data, governance, and user workflow match the operating need.

What still has to be solved

A tested transformation can be technically correct while the source fact, entity match, business rule, or time assumption remains wrong.

Model names and columns do not automatically communicate evidence class, confidence, legal meaning, or the decision a human is authorized to make.

A platform can be an excellent system of record or execution and still leave an intelligence problem. Fields, messages, meetings, campaign events, listings, and transactions do not interpret themselves. Entity identity, evidence quality, time, ownership, and commercial meaning still have to be resolved.

This is not a claim that the vendor has failed. It is the boundary between buying software and operating an accountable intelligence system. Every organization must still define its entities, event semantics, source precedence, decision rights, time rules, and correction path.

The vertical-agnostic intelligence overlay

Delta Arc defines the evidence and decision contracts that dbt models must preserve and test, including original values, observation times, conflicts, and confidence boundaries.

The combination creates reproducible transformations without confusing a derived table with proof beyond its sources.

Delta Arc is the vertical-agnostic interpretation layer across those systems. It does not require an organization to replace the applications people already use. It joins the events, preserves provenance, distinguishes activity from state change, and returns the next useful decision to the CRM, inbox, collaboration channel, report, or public product.

Where it fits—and what to validate

Delta Arc can work with the selected platform rather than prescribing a replacement. The architecture begins with the decision and evidence model, then uses the existing CRM, collaboration, data, marketplace, search, or communication product as the appropriate system of record or execution.