The concepts, metrics, and frameworks behind building software when AI writes most of the code. Each term is defined in one place, in plain language.
The AI-Native SDLC Maturity Model is a five-level framework (L0–L5) describing how deeply an engineering organization ha…
AI-native software delivery is a way of building software in which humans express intent and AI agents do most of the im…
Engineering memory is an organization's persistent, queryable record of how and why its software was built — a graph con…
An execution manifest is the complete, self-contained context payload attached to a work order — the intent, constraints…
The Intent-to-Production Thread is unbroken traceability from a unit of business intent to the exact commit that shipped…
In software delivery, a landmine is a latent architectural risk captured on a decision record before it detonates — a kn…
Release confidence is a composite readiness score that answers 'should we ship this?' by combining signals such as test …
Reverse-engineering onboarding is the practice of pointing an AI-native delivery platform at an existing repository and …
Spec drift is the gradual, usually silent divergence between what a specification says a system should do and what the c…
A system of record for software delivery is the authoritative, governed store of an organization's intent, decisions, wo…
The Orchestration Gap is the widening distance between how fast AI can generate code and how slowly organizations can pl…
A work order is a self-contained, agent-executable unit of software work: a scoped task compiled from business and techn…
Agent orchestration is the coordination of multiple AI agents — assigning work, supplying context, sequencing steps, app…
Approval gates are governance checkpoints in a software delivery pipeline that a change must pass before it can progress…
An architecture decision record (ADR) is a short document that captures a significant technical decision, its context, a…
In software delivery, a change's blast radius is the full set of components, contracts, services, and teams it can affec…
DORA metrics are four widely-used measures of software delivery performance — deployment frequency, lead time for change…
An engineering knowledge graph is a structured network of an organization's software artifacts and their real relationsh…
MCP orchestration is the practice of coordinating AI coding agents through the Model Context Protocol (MCP) — a standard…
Shadow AI in the SDLC is the unmanaged, unofficial use of AI tools and coding agents by developers — outside any central…
A software factory is an environment that standardizes and automates how software is produced end to end — from intent t…
Spec-driven development is an approach in which a clear specification — requirements, acceptance criteria, and constrain…
Vibe coding is an informal style of AI-assisted development in which a person prompts an AI agent in natural language an…