Glossary
A shared vocabulary keeps the whole platform speaking one language. Each entry below states what the term is and where it appears.
Products
Nucleus
The connect product. Brings third-party software onto the OneOps Gateway as one live mesh: import, harmonize, and synchronize records, provision and deprovision business assets, and stream operational data into the OneOps Datalake. See Nucleus.
Platform
The shared substrate the products run on: identity and access, tenancy, the connector catalog, audit, usage, and the workspace shell. Platform is infrastructure, not a product you buy on its own. The product that connects your systems to it is Nucleus.
AI Studio
The create surface: create and test skills and agents, plus chat, AI-run workflows, projects, and the drafting editor. Formerly named AI Workspace, then Agentic Studio. See AI Studio.
AI Operations
The run surface: the runtime quality environment where skills and agents operate full time, with protections against AI harm, AI cost, and protected-data exposure, plus logging, emergency stop, undo, and monitoring. See AI Operations.
AI Marketplace
The place to find and buy skills, agents, and knowledge, built by the OneOps team today, tested in AI Studio and operated in AI Operations. Publishing by other organizations is planned and will be gated on a quality and security review model. See AI Marketplace.
Core concepts
Tenant
An isolated workspace with its own data, members, and settings. Everything a user sees is scoped to the tenant they are signed in to. Tenants never read each other's records, which is the boundary the rest of the platform is built around.
Connector
A configured link to an external system. A connector wraps another service's API as a set of typed operations, so the same actions look identical no matter which underlying system fulfills them.
OneOps Gateway
The single seam every connected system reaches the platform through. Connecting a system to the gateway is what Nucleus does, and it is why the same operation looks identical no matter which underlying system fulfills it.
OneOps Datalake
The landed store of operational records Nucleus streams in, so reporting and analysis run over a consistent history rather than over live source calls. Read-only to the products that consume it.
Catalog
The browsable list of connectors available to a tenant. Adding a connector from the catalog is the onboarding step that brings an external system online. See Integrations.
Sync
A scheduled or on-demand pull that brings external records into the platform and keeps them current. A sync has a cadence, a window, and a record of its last successful run. See Sync and schedules.
Workflow
An ordered set of steps that accomplishes one operational outcome. In AI Studio, the assistant can run multi-step work and hand parts to specialist agents.
Scenario
A named situation paired with the response it calls for: a trigger, the steps to take, and the signal that confirms resolution.
Audit log
An append-only record of who did what and when, scoped to your workspace. It records that a record was accessed, never the value, and is never edited in place. See Audit and usage.
Role
A named bundle of permissions assigned to a member. Roles decide which surfaces appear and which actions are allowed, and they map to least privilege. See Roles and access.
Alert
A signal that a metric crossed a threshold or a process failed. Alerts are designed to be actionable, each pointing at how to respond. Alerts appear on the intelligence dashboards' dispatch and operations boards.
AI Marketplace offerings
Compliance Link
A first-party knowledge offering available through AI Marketplace. It provides cited, tier-ranked compliance and regulatory intelligence plus provider exclusion screening. It is not a separate product or a chat surface. See Compliance Link.
Intelligence dashboards
The cross-system 360 operations view: dashboards, a fleet map, a dispatch board, revenue, and utilization. A first-party item you add from AI Marketplace and operate through AI Operations. See Intelligence dashboards.
Services
Support
Included AI support, with budget for escalation to a person when it is needed.
Consulting
Expert time, scoped and sold under a statement of work.
AI Studio terms
Agent
A named specialist the assistant can put to work, each with its own role, model, and tools. An agent can carry several skills plus an overarching workflow, which is what distinguishes it from a single skill. A workspace starts with a stock roster and a tenant admin can edit, duplicate, add, or disable agents from there.
Artifact
A richer piece of output an assistant produces (a table, code, a diagram, a map, a document, a slide deck, an image), shown as a card that opens in a side panel and can be saved to a project's Artifacts board.
Delegator
The mode that fans one request out to several specialist agents and folds their results back into a single answer, showing each delegated step as it starts and finishes.
Project
A grouping of related chats under shared instructions and reference files, with tabs for tasks, memory, files, notes, skills, and pinned artifacts. See Projects.
Project memory
The visible, editable record of what the assistant has learned about a project: standing facts it carries into every chat in that project rather than being told again. See Projects.
Skill
A discrete, reusable capability an agent can run: a set of instructions, the reference material bundled with them, and the tools the skill is allowed to use. Skills live in a workspace catalog and can be pinned to a project. See Skills.
Exemplar artifact
A real sample of a skill's output, rendered in the skill's detail panel so you can see what the skill produces before you run it. See Skills.
Compliance terms
Evidence tier
The authority ranking on a regulatory fact in Compliance Link. Tiers one through three are authority (federal sources and state configuration); tiers four and five are context, never authority. See Compliance and screening.