Set your modeling conventions

Your modeling conventions are the organization-wide decisions that shape every model: which categories classify your business entities, how technical names are derived from business names, and which methodology names your data layers. Categories and the naming convention live together under Settings → Conventions. The methodology editor and abbreviation rules live at Settings → Preferences. This guide explains each area and how to manage it.

To open your categories and naming convention, go to Settings in the sidebar and select Conventions. The section has three areas: Categories and Naming convention, covered in turn below, and Classifications, covered in Classification settings. Methodology and abbreviation rules are configured separately at Settings → Preferences and are described later in this guide.

Choose your categories

The Categories area is where your organization decides how it classifies business entities. You keep one active set of categories, and every entity in your model is tagged with one of them. The area shows a single Define your categories button and, once a set exists, a read-only Your categories table.

Each category in your set carries three things beyond its name:

a Type, either Event (something that happens, such as an Order or a Payment) or Concept (something that simply is, such as a Customer or a Product),
a W, the single interrogative the category answers: who, what, where, when, why, or how,
an AI context, a short note in your own words describing what the category means. The business-modeling and glossary AI read this note when they analyze your input, so their suggestions match how your organization thinks.
A category is a named kind of business object, for example Individual Person, Order, or Location. Each entity in your model is tagged with one.
An archetype is the broad family a category belongs to, for example Event, Party, or Place. Archetypes keep a long list of categories organized and readable. DeltaVault supports two modeling approaches, and the exact set of archetypes depends on the approach your catalog follows (see Choosing an approach below).
Your category set is the categories your organization works with. You keep one active set, and saving the wizard replaces it.
The set always includes an Other category, a catch-all for entities that do not fit any named category, so nothing is ever left unclassifiable.

You do not invent categories from scratch. DeltaVault ships a curated library of categories, grouped into the archetypes of the approach you pick, each with a short starting definition you can adopt and adjust.

Define your categories with the guided wizard

Select Define your categories to open the wizard. It has three steps, and nothing is saved until you finish.

Pick a starting approach. Choose Ensemble Logical (Event / Person / Place / Thing / Other) or Nine-Pillar (Party & Party Role tradition) (compared in Choosing an approach below). The approach sets which archetypes organize your categories. Selecting an approach reveals the industry choice.
Pick your industry. Choose Baseline (no industry), which keeps the approach’s standard wording, or one of the five industries, which re-skins the archetype names and AI guidance for that industry (see Choose an industry below). Baseline is preselected, so you can continue without choosing one. Select Use this approach to continue, or Back to change the approach.
Edit your categories. The categories for your chosen approach and industry open in an editor. Each category arrives with its AI context pre-filled from the library definition, so you have an editable starting point. For each one you set its Type (Event or Concept), its W (the interrogative it answers), and its AI context, and you can rename, reorder, remove, or add categories from the library. The AI-context field offers AI drafting in place: see Use AI-assisted fields. The Other catch-all is always kept. Select Save to make this your active set.

To start over from a different approach or industry, select Re-prime on the edit step, which returns you to the approach picker.

Choosing an approach

The two approaches draw from the same library of categories but group them into different archetype families, so the one you pick changes how your categories are organized and labelled, not which categories you can use.

Ensemble Logical uses a coarse set of five archetypes: Event, Party, Place, Thing, and Other. It is the simpler starting point when you want a small, stable set of families and do not need to distinguish, for example, a product from an agreement at the archetype level.
Nine-Pillar uses a finer set of ten archetypes: Party, Party Role, Product or Service, Agreement, Event or Transaction, Place or Geography, Resource or Asset, Classification or Reference, Document or Communication, and Other. It splits things and concepts into more specific families, which suits organizations that want their roles, offerings, agreements, and reference data called out distinctly.

Both approaches always include the Other catch-all, and you can switch the framing later by running the wizard again. The choice affects archetype labels and the AI context the categories carry; it does not lock you out of any category.

Choose an industry

After you pick an approach, the wizard asks you to pick an industry on its own step. The choices are:

Baseline (no industry)
Regulated Finance and Insurance
Retail, Consumer Goods and E-commerce
Manufacturing, Supply Chain and Logistics
Healthcare, Life Sciences and Human Services
Energy, Utilities and Telecommunications

An industry re-skins the archetype names and the AI guidance for your industry. It does not change which categories are available, how the archetypes are structured, or how entities are classified. The category mapping is identical with or without an industry.

Baseline (no industry) is preselected and keeps the approach’s standard wording. Choosing an industry means the AI uses industry-specific terminology when it helps classify and model your data for that catalog. To change it later, run the wizard again from the approach picker and pick a different industry.

Change your categories later

Select Define your categories again to run the wizard over your active set: add or remove categories, reorder them, or change their type, interrogative, or AI context, and save. Changing your categories affects how entities are classified going forward. Entities you have already classified keep their current categories until you change them.

Name technical names from business names

Every entity and attribute has two names: a business name (how people in your organization talk about it) and a technical name (how it appears in generated code and physical tables). The Naming convention area sets a single rule that derives the technical name from the business name, so the two stay consistent without anyone typing each technical name by hand.

You set the convention separately for entities and for attributes. Each has two casing selects, Name for the business name and Technical name for the derived name, for example PascalCase for entity technical names or snake_case for attribute technical names. The abbreviation rules configured under Settings → Preferences (for example always shorten “Number” to “Nbr”) apply as the technical name is derived, and the Max identifier length in the Shared section caps how long it can be.

A live preview shows how a sample business name comes out as a technical name while you adjust the settings, so you can see the effect before you save.

While you model, the technical name follows the business name for you. When you give an entity or attribute a business name, its technical name fills in from the convention, and it keeps following as you rename. As soon as you edit the technical name yourself, it stops following and keeps exactly what you typed, so a name you set on purpose is never overwritten by a later rename. The same convention is used everywhere a technical name is derived, including the “generate technical name” action on an entity, so names read consistently across the whole model.

Choose your methodology

The methodology editor lives at Settings → Preferences, a section only administrators see. To reach it, go to Settings in the navigation and select Preferences; the same section carries your Preferred language, Modeling persona, and Abbreviation rules. Your methodology determines how the functional data layers are labelled (the names used for layers such as Bronze, Silver, and Gold) wherever they appear in the model and in generated code. Pick the methodology that matches how your team works; if you use your own name for a layer, you can override an individual label. The choice is organization wide, so everyone sees the same layer names.

Your conventions shape AI recommendations

The conventions you define here are injected into the AI skills that model on your behalf, so recommendations arrive already speaking your language:

Categories drive event and concept recognition everywhere the AI proposes or reviews model structure: entity recommendations, relationship suggestions, the business modeling workshop’s discovery and judge passes, relationship discovery, and the model review’s category conformance dimension all read your lanes, each lane’s nature and question, and your guidance text.
Naming convention shapes proposed names: workshop discovery and entity enrichment name every proposal to match your entity and attribute casing.
Classifications bound AI classification suggestions: column classification and classification propagation only propose values from your schemes, including custom schemes you define.

Administrators can see and adjust exactly which skills read which convention in the AI Playground’s Context tab; see Manage AI skills in the Playground.

Lock your conventions

By default, anyone with edit permission can change these conventions. An administrator can lock the Conventions section so that only administrators make changes and everyone else sees it read-only. Use Lock your categories, beside the Categories heading in Settings → Conventions (it reads Unlock your categories while locked); it covers the categories and the naming convention there, and everyone else sees “Locked by an administrator. Read-only.” The same lock also governs the methodology editor at Settings → Preferences: even though the methodology editor lives there, the Conventions lock controls whether anyone can edit it.

The lock is an organization-wide setting and is separate from branch read-only: it does not stop normal modeling work, only changes to the conventions themselves. While locked, editors and viewers can still read the conventions and see how names derive, but the controls stay disabled until an administrator unlocks. The default is unlocked.

Where to go next

Tag entities with their categories on the canvas: see Business entities.
See how modeling, the glossary, and governance fit together: see Business modeling and governance overview.
Draft category AI context and other prose with AI help: see Use AI-assisted fields.