Why "Substrate"?
"Substrate" is used in several fields. The meaning differs in each, but the idea is the same. In biology and chemistry, a substrate is the molecule an enzyme acts on; without it the enzyme does nothing. In geology, it is the rock or sediment under the soil, which determines what soil and life can develop above it. In materials science and microelectronics, it is the layer everything else is deposited on: silicon under the transistors, sapphire under the LED. In linguistics, a substrate language is the indigenous language whose grammar persists in a colonising language and shapes it. In fermentation and cooking, it is the medium the microbes grow on.
In every case, a substrate is the base underneath, often unseen, that determines what can grow, happen or be built on it. That is the relationship between what a Map engagement writes down and what gets built afterwards.
For Orbital, a substrate is a firm's captured expertise. It has three layers, which are what the firm already knows: quantitative, subjective and procedural. Three wrappers are defined from the layers: question archetypes, constraints and safety, and the grounding constraint, which is what every output is tied to. The Document Substrate is the software that applies a firm's substrate to client documents.
The word says six things:
It names the right layer. The Map phase does not record "what the business does". That is the visible work. It records the knowledge the visible work depends on. Tax statute under tax planning. The PMC under cycling coaching. Judicial precedent under legal advice. Building code under construction. The substrate is not the work. It is what the work depends on.
It implies stability. Substrates change slowly. Tax law, fitness science, judicial precedent and regulatory frameworks evolve, but not quickly. AI models change every six months; substrates change over a decade or longer. Calling this layer the substrate says it is the long-term asset of the engagement: the part that does not need redesigning when the tooling changes. Different things can be built on a good substrate over time: a web app and an MCP server, this year's product and next year's.
It implies sequencing. Substrates come first. Soil does not form without rock under it, and chips are not fabricated without a wafer. Calling the Map output a "substrate" shows that the Map → Architect → Build order is necessary, not optional. Skipping the Map is not a shortcut. The architecture would be designed with nothing to base it on.
It implies what is possible and what is not. A silicon substrate supports certain circuits; sapphire supports others. A clay subgrade supports certain foundations; granite supports others. The substrate determines what can be built on it. The same applies here: the Map output is not a wish list of features. It is a set of constraints the architecture has to respect. Domain knowledge determines what the AI system can do, and the word says so.
It credits what was already there. A substrate accumulates: tax law over centuries, judicial precedent over a millennium, fitness science over fifty years, a partner's deal patterns over a career. Calling the Map output a "substrate" acknowledges that the consultant did not invent the knowledge. It was already in the people and the firm. The Map phase writes it down. This matters because the engagement records value the firm already has, and does not impose a system from outside. Experienced practitioners prefer to be approached that way.
It separates the business-specific from the portable. The architecture is largely portable across domains. CTI's three-tier memory, hybrid retrieval and layered prompts work for accounting too, as the worked examples across six professions show. The substrate is the part that is not portable. It is specific to the business, it cannot be replaced, and it is the firm's competitive advantage.
Each alternative loses something. "Knowledge base" implies a database, which is one particular implementation. "Domain model" implies a software model and a developer audience. "Expertise" implies humans. That is where the knowledge is now, but not where it needs to end up. "Ontology" is too academic. "Foundation" is close, but generic and overused in consulting. "Context" is too vague: every prompt has context. "Substrate" is precise enough to mean something concrete, and it says nothing about implementation: it could be prose, vectors, structured data, or all three. It also carries the order of Map → Architect → Build.
The word also argues against "AI as plug-and-play". Generic AI assumes no substrate is needed, or that the model is the substrate. The word names the missing layer and explains why generic AI produces generic output: the model has nothing specific to the firm to reason from. The rest of this website makes the same argument at length.
In short, the word tells you what kind of layer this is, where it sits, how it relates to what is built on it, why it has to come first, why it stays stable while the tools change, and why it is the part that is yours.
Worked examples across six professions · Map → Architect → Build, in practice · Read the CTI case study