Glossary

Plain-English definitions of every Hub-specific term you'll see in the user documentation.

Access Request

A formal request to view or interact with restricted data. If a field is masked or a space is restricted, you can submit an access request explaining why you need the data. An approver reviews it and may grant temporary access.

Agent

An AI automation worker that runs inside your Space. An Agent responds to triggers (messages in Ask Anything, schedules, webhooks, events) and takes action: creating tasks, updating records, leaving comments, building dashboards, sending notifications. See Agents.

Artifact

A substantial piece of content an agent produces inside a Conversation — a document, a spreadsheet, some code, or an interactive page. Artifacts appear as cards in the conversation and can be opened in a split view so you can read them full-size without losing your place.

Ask Anything

The conversational AI assistant docked in a side panel on every page of your Space. You open it with the purple Ask Anything button at the top right and chat with the assistant about whatever you're looking at. The assistant uses the page you're on as context, suggests prompts tuned to that page, and saves every chat as a Conversation. See Ask Anything.

Appearance

The visual look of your Space: logo, accent colour, font family, spacing, and light/dark mode. Administrators can customize the appearance to match the organization's branding.

Attribute

A field or property on a record. Custom attributes are user-defined fields added by operators without code — for example, "Risk Level" on a Task or "Funding Round" on a Company.

Chat

The conversation surface for talking to Agents — the panel that opens when you click Ask Anything. Your messages and the agent's replies are stored as Conversations you can return to later. See Ask Anything.

Blueprint

A reusable bundle that defines a kind of product: one Process plus the attribute schemas for the Products, Subscriptions, and Payment Line Items that use it. Operators create Blueprints in the Tenet admin space; every Product is created against a Blueprint and inherits its process and field set. See Blueprints.

Breadcrumbs

The trail of links across the top of every page that shows where you are and how you got there — for example, Explore › Users › Jane Doe on a user detail page. Each step is clickable and takes you back to that level. Hub fills in the trail automatically based on the page you're on.

Chat Skill

A tool or capability that a platform operator curates and makes available to users in Ask Anything. Chat Skills help users ask the assistant to do specific things — for example, "generate a report" or "create bulk tasks". Operators define which Chat Skills are available in each space. See Chat Skills.

Comment

A note or message on a task or record. Comments support markdown, file attachments, and @mentions. You can edit or delete your comments.

Conversation

A single thread of back-and-forth messages between you and an agent in Ask Anything. Conversations are saved automatically with a short auto-generated title and live in your saved conversations list. Scheduled runs of a saved prompt each produce their own conversation.

Custom Attribute

See Attribute.

Dashboard

A customizable view of metrics and data from your Space. A Dashboard contains Widgets — stat cards, charts, lists — that you or your administrator can add, remove, or rearrange.

Data Table

A structured list of records (employees, companies, tasks, trades) organized in columns with sorting, filtering, and bulk actions. Data tables can export to CSV or Excel.

Derivation

A computed field calculated from other attributes. For example, a derivation might calculate "Total Held Value" from holdings, or "Days Since Created" from the creation date. Derivations are updated automatically when source attributes change.

Dynamic Resource

A user-defined record type (like a custom table). An operator creates a Dynamic Resource by naming it, adding columns (custom attributes), and Hub instantly provides a CRUD interface and agent tools. Records in a Dynamic Resource are stored and queryable like built-in records.

Event

A real-world happening that triggers automation. Examples: "Company Created", "Employee Added", "Task Closed". Agents listen for events and run skills in response.

HR Provider

The payroll or human-resources system your company already uses to keep employees, contracts, and pay records — for example Gusto, Rippling, ADP, BambooHR, or Workday. A Company admin can connect an HR provider to a Company Space so the employee directory stays in step automatically. See HR Provider Sync.

Execution

A single run of an Agent. When a skill is triggered, it starts an execution. The execution has a status (running, waiting, completed, errored), a conversation history, and results. You can view past executions in the agent log.

Knowledge Base

A shared library inside a Space for reference material. Has two halves: Documents (files you upload) and Wiki (pages you write directly in Hub, organised in a tree). Available in Initiative, Platform, Support, and Tenet workspaces. See Knowledge Base.

Document (Knowledge Base)

An uploaded file in a Space's Knowledge Base — a PDF, slide deck, contract, image, or anything else up to 50 MB. Has a title, an optional description, and a download link. Distinct from the smaller files you attach to a specific task or record (those are File Attachments).

Wiki Page

A page you write directly inside Hub's Knowledge Base, with a Markdown editor and AI assistance. Wiki pages live in a tree of parent and child pages, keep a version history, and are visible to every member of the Space.

MCP (Model Context Protocol)

The open standard Hub uses to let AI assistants (Claude Desktop, Claude Code, ChatGPT, and others) connect securely to your workspace. When you configure an AI assistant with Hub's MCP server URL, it can read data from and take actions in your Space on your behalf — authenticated through OAuth. See Connecting AI Assistants.

Lifestyle Strategy

A glide-path investment strategy for workplace pensions. The strategy is divided into bands by years to the target retirement date; each band defines a target mix of funds. As an enrolled employee approaches their target date, Hub rebalances their portfolio across bands automatically. See Lifestyle Strategies.

Member

A person who belongs to a Space with a specific role (Owner, Editor, Viewer, etc.). Each Member has permissions appropriate to their role.

Mention

A reference to another person using @name in a comment or task description. Mentions are highlighted and trigger notifications to the mentioned person.

Glide Path

The path a member's pension portfolio takes from a higher-growth fund mix when retirement is far away to a more conservative mix as retirement approaches. A Lifestyle Strategy defines this path as a sequence of bands.

Notification

An alert sent to you when something happens in Hub: a task is assigned to you, someone mentions you, an agent finishes. You receive notifications through channels you've opted into (email, in-app bell, phone push, browser push). The in-app bell opens a side panel listing your twenty most recent notifications. See Notifications.

Notification Channel

A way to receive notifications: email, in-app bell, mobile push, web push. You choose which channels you want in your notification settings.

Ontology

The structure of your data: custom attributes, derivations, and the relationships between record types. Platform operators maintain the ontology in the Tenet space.

Platform

The Hub software itself and the central admin workspace where operators configure themes, capabilities, and global settings.

Project

A named container that groups related tasks under a single banner with a start date, an optional end date, a description, and any custom fields. Projects live in Initiative, Platform, Support, and Tenet workspaces (when Project Management is enabled), and have Active and Closed tabs. See Projects.

Preferences

Your personal settings — appearance mode (light/dark/system), date format, language, and currency — opened from the sliders icon in the top right. Preferences are per-user and follow you across every Space. See Preferences.

Feedback

A short form opened from the chat bubble icon in the top right where you can send a bug report, an idea, or a question to the team running your platform. See Feedback.

Saved Prompt

A reusable prompt you've saved from a message in Ask Anything. Saved prompts are yours — only you see them. You can fire a saved prompt into any chat as a starting point, or set it to run on its own (see Scheduled Prompts).

Scheduled Prompt

A saved prompt set to run on a repeating schedule — daily, weekly, or monthly. Each run opens a new Conversation, executes the prompt, and optionally emails you the result.

Source Path

The dotted location of a value inside an external provider's payload, written into a custom attribute so Hub knows where to copy that value during a sync. For example, a Source Path of individual.dob on a "Date of Birth" attribute tells Hub to read the date of birth from the HR provider's individual record. The shortcuts _preferred_email and _preferred_phone pick the best work contact the provider returns.

Skill

The trigger configuration for an Agent. One Agent can have multiple Skills:

Chat — triggered by you sending a message in Ask Anything
Scheduled — runs on a saved prompt's schedule (see Scheduled Prompts)
Webhook — triggered by an external system via HTTP
Event — triggered by a domain event (e.g., company creation)

Space

A self-contained workspace for a specific context — a company, an advisory group, a project team, or your personal workspace. Each Space has its own members, tasks, data, and settings. You can belong to many Spaces and switch between them.

System Prompt

A templated prompt that a platform operator creates and surfaces to users in the Ask Anything chat composer. When users open Ask Anything, they see suggestion chips for relevant System Prompts on that page — for example, "Summarise open tasks" on a task list. System Prompts help users quickly get helpful answers without typing from scratch. See System Prompts.

Task

A unit of work: something that needs to be done. Tasks have a title, description, status (open or closed), assignee, optional reviewer, due date, and can have comments and attached documents. See Tasks.

Task Automation

Automatic task creation triggered by domain events. For example, when you create a company, Hub automatically creates onboarding tasks. When a task is closed, related tasks may also close.

Tenant

A separate Hub environment. Each tenant has its own web address and keeps its data completely isolated from every other tenant — members, companies, tasks, documents, and settings don't cross between them. User accounts don't either: signing in to one tenant doesn't sign you in anywhere else. Tenet admins who work across more than one tenant can hop between them from the Switch tenant entry in the user menu (see Operator Spaces).

Tenet

The admin workspace where platform operators configure Hub: enable/disable features, create Agents, define custom attributes, customize themes, and manage ontology.

Theme

The visual appearance configuration of a Space or the platform: colour scheme, font family, border radius, spacing, appearance mode (light/dark). Administrators can choose from preset themes or create custom ones.

Widget

A single component on a Dashboard: a stat card, chart, table, or other visualization. You can add, remove, and rearrange widgets to customize your dashboard.

Verification

A compliance record showing a member or company has completed a particular check — KYC, AML, proof of address, etc. Each verification has a type, a status (Pending, In Review, Approved, Rejected), an optional expiry date, and any documents collected. See Verifications.

Webhook

An HTTP endpoint exposed by Hub that external systems can POST to in order to trigger an Agent skill. Webhook skills let external applications (CRMs, accounting systems, etc.) integrate with Hub.

Process

A visual diagram administrators build to automate work. A process is made of a trigger, a sequence of AI tools, and flow-control nodes (conditional, foreach, switch, try/catch, parallel) connected by arrows. When the trigger fires, Hub runs the diagram. Processes come in three types — pipeline, product, and fees. See Processes.

Pipeline

A type of Process used for general automation: respond to events, run on a schedule, accept webhooks, or react to member actions. Most agents are pipelines.

Product (Process type)

A type of Process that defines a financial product offered to members — its rules, eligibility, and contribution behaviour. Product processes live inside a Blueprint.

Fees (Process type)

A type of Process that calculates fees for a member or transaction.

Trigger

The starting point of a process — what makes it run. Triggers can be events, schedules (cron), webhooks, member actions, or payroll dates.

Getting Started — learn the basics
What is Hub? — broader context