Glossary

The words Emargen uses, as the app spells them, each in a line. Follow a word to its page for the whole idea.

A

Academy: where you learn what your organisation knows, at your own pace. It maps its systems, roles and terms.
Admin: where administrators set up and run the organisation. Its pages are the setup’s steps, kept after the launch.
Affordances: the pieces a decomposition names. They run from what people see and do to the code and data underneath.
Agent: the AI in a chat beside most pages. It reads what your organisation has connected and works on the page with you.
Agent access: what people’s own agents, like Claude Code, may do in the organisation as them.
Agent Notes: the notes your agents leave you, for you alone. Coding agents leave them as they work, and any agent leaves one when you ask.
Agent task: a job you hand the agent on a schedule. It writes a topic, fills a list, or keeps a topic up to date.
AI Overview: the agent’s reading of what was decided. It covers one place or the whole Decision Map.
Analysis: what Emargen writes about a repository’s code, part by part, in words the whole team reads.
App: a tool your team needs that Emargen doesn’t have, which Emargen’s agents build from a brief you shape. A small program inside a topic is an app too.
App builder: where you start an app, shape its brief and follow its builds. It’s under Workspace.
Approach: one way to build an initiative, laid out in parts and checked against every requirement.
Approval: a published initiative’s yes, given on its Kickoff by someone allowed to approve. Emargen keeps who gave it and when.
Approval gate: an initiative’s setting that holds it from Approved until someone other than its owner approves it.
Archived: a topic set aside as out of date.
Arranges: someone who may put topics on a list kept by hand and order them. They take topics off it too.
Autonomous: the pace at which an agent goes through a checklist on its own, stopping only for the steps that wait on your answer.

B

Block: one piece of a topic’s section. It can be a paragraph, a table, a code check or a decision.
Blockers: what holds a slice or a task up, and what it holds up in turn, drawn as a map. A blocked task can’t be claimed.
Board: a saved way of looking at topics or initiatives. You choose conditions or the lists you keep.
Brief: what an initiative’s problem is, what was decided about it, and how things work today.

C

Calendar: what’s due and what happened. It also shows what the agent does for you on a schedule, day by day.
Call word: a persona’s short name, like @security, which asks that persona to review an answer. A skill’s starts with /, like /reply.
Card: one question in practice, asking one fact. In a chat, a card holds a change for you to confirm.
Challenge: the same few cards for you and a colleague, each answered on your own.
Change request: a request to change what an approved initiative shaped, set out as options for its owner to decide.
Chat: your conversation with an agent beside a page. Nobody else reads it.
Check: a block where the agent checks a question against your code and sources. It answers Exists, Exists in part, Missing or Unclear.
Checklist: a job of several changes an agent does for you, written down as steps you start once.
Checks: the checks your workspace adds to every initiative’s Kickoff.
Coding agent: Claude Code, Codex or another agent on an engineer’s computer, connected to Emargen as them. It builds tasks, writes topics and leaves notes.
Comment: what people write to each other on a topic or an initiative, apart from their chats with the agent.
Company role: a role your company’s own sign-in carries for a person, which can give them Emargen roles.
Connected assistant: Claude or ChatGPT, signed in to Emargen as you from its own app. It’s listed on your profile’s Security.
Connection: a tool your organisation’s knowledge lives in, like Jira or Confluence. Once connected, the agent can search it. The agent uses one sign-in for everyone.
Core goal: the one requirement an initiative exists for.

D

Decision: a choice a person makes on a topic, an initiative or the Landscape. It’s kept with who took it and why.
Decision Map: what was decided, where and when, beside what’s still open.
Decisions: the choices waiting on you to decide or approve, from every topic and initiative. You decide each one where you read it.
Decomposition: the agent’s map of how an initiative’s chosen approach would be built in your code.
Default: the model a kind of work runs on, chosen by whoever manages models.
Delivery: the stage when an approved initiative is being built, slice by slice.
Document: the part of a topic the agent writes, in sections. It’s built from what you said and what it found.

E

Edits: on an initiative, someone who shapes it as the owner does. On a board, a list, a template or a persona, someone who may change it.
Everyone: shared with everyone in the organisation.
Everything: the board every area has, with no conditions. It’s the widest choice of what a token or an assistant may do.

F

Fact: one thing to know about a spot, asked by its cards.
Field: a short fact a topic carries, like the customer or the day it’s due. Its template says which fields it has.
Fit check: every requirement of an initiative set against every approach, one requirement at a time.
Flow: how a kind of work goes, step by step, and who does each step. Each case goes through it as a run.

G

Guided: the pace at which an agent waits for your Next at each step, so you can read why.

I

Inbox: one stream of everything that arrived for you. What you have to reply to and do sits at the top.
Initiative: work that should be built, shaped with the agent into a plan someone approves. It’s followed while it’s built.

K

Key: your organisation’s letters and a number, like ACM-281. They name a topic, an initiative or a piece of work. In what people write, a key shows alone, with its title on hover.
Kickoff: one page for the people who approve an initiative and those who build it. It says what the initiative will do and how.
Kit: a ready-made connection for a tool like Jira Cloud or Confluence Cloud. It carries what to search and how to check it.

L

Label: a word for what work is about, like a customer’s name, that anyone puts on topics and initiatives.
Lane: one row of the Decision Map, for each place decisions were taken.
Landscape: your organisation’s own names for its systems, people, roles and terms, which the agents read.
Launch: opening a newly set-up organisation to everyone in it.
List: topics or initiatives collected under one name, by hand, by conditions or by an agent task.
Live part: a block that shows work as it stands each time the page opens. The work comes from Emargen or a connected tool.
Look-back: a coding agent’s reading of your earlier sessions on your computer, kept as notes.

M

Model: the AI a kind of work runs on, bought from one of your organisation’s providers.

N

Note: one thing an agent told you, for you alone. It may be a decision it took, a todo, a trap or an idea.

O

Only me: the sharing choice that keeps a topic, an initiative or a board to its owner. Everything new starts this way, except a subtopic.
Organisation: your company in Emargen, with its people, its AI, its knowledge and all their work.
Owner: the person a topic or an initiative belongs to, who alone changes who sees it.

P

Persona: how the agent approaches your work in a chat. You pick it, with its instructions and its model.
Personal access token: a key you make on your profile for a coding agent. It says what the agent may do and for how long.
Pin: you keep an item at the top of your Inbox.
Place: in a decomposition, the screen or mode the pieces live in. A portal page or an email is one. On the Decision Map, where decisions were taken.
Plan: how much Emargen runs for your organisation at once. You see it on Plan, under Admin.
Practice: a few cards a day from your tracks, each fact coming back when you’re about to forget it.
Proposal: a change to the Landscape someone suggests, waiting for someone allowed to approve it.
Provider: a company your organisation buys AI from, like OpenAI or Anthropic.
Publish: keeps an initiative as it stands and sends it for approval. Coding agents read what was published.

Q

Questions: in a decomposition, what the agent couldn’t decide from the code, and what holds meanwhile.

R

Reads: may open a topic or an initiative and comment. Their own chat explains it and drafts comments, and changes nothing.
Recipe: what a repository’s scan recognises, and what its analysis looks for.
Record of AI use: a line for every call to your organisation’s models, with what it cost.
Repository: your organisation’s code, added once so the agent can read it.
Requirement: what must be true when an initiative is done, with a level that says how much it matters.
Role: what a person may do beyond their own work, like inviting people or approving initiatives. Sharing decides who sees an item, not a role.
Run: one case going through a flow, with its own key. You see them and start them on Runs, in the menu.
Runs: on a flow, someone who may start runs of it.

S

Scan: what Emargen does first with a repository. It finds the areas the code divides into and proposes recipes for each.
Searchable content: the part of a connection the agent may search.
Section: a part of a topic’s document, with a heading, filled with blocks.
Session: the stretch of work a coding agent did with you in a repository. Its notes are read together.
Setup: the steps a new organisation goes through before its launch.
Shaping: the stage in which an initiative is worked out with the agent, before it’s published.
Shared: seen by the people you name.
Shared with you: the group of your topics list that holds topics shared with you by name, until one is on a list.
Sharing: who can open a topic or an initiative, and what each person may do.
Skill: know-how the agent loads when the work calls for it.
Slice: one step of an initiative you can show working on its own.
Sources: everything an initiative was worked out from.
Spike: a short investigation into something an approach depends on. It looks in the code, your knowledge, the record or the web.
Spot: one thing to learn, with its facts and the cards that practise them.
Star: your own mark on a topic or an initiative, which keeps it under Starred in the menu.
Start here: Emargen’s own part of the Academy map, which shows how work happens in Emargen.
Subtopic: a topic under another topic. At first, whoever sees the topic above it sees it too.
Suggested answer: the agent’s draft answer to something that waits on you, which you use or change.

T

Takes part: on a topic, has their own chat with the agent. They may change the topic through it.
Task: one piece of a slice’s work, which a coding agent builds. An agent task is a job on a schedule, and a task in your Inbox is something to do.
Template: the shape of one kind of topic, its sections and its fields.
To do: the group at the top of your Inbox with Emargen’s tasks for you and your notes’ open todos.
To reply: the group at the top of your Inbox with the comments that tag you and the questions waiting on your answer.
Topic: where you work something through with the agent, into a document others can read. It can start from research, a plan or a customer’s question.
Track: your own collection of things to learn from the Landscape, framed with the Academy agent.
Triage: the group of your topics list that holds your own topics, until one is on a list.

U

Uses: on a template, persona or skill shared with you, you may use it. You can’t change it.

V

Version: a topic as it read after one sitting of work, with who changed what.
Views: on a board, a list or a flow, someone who may see it and nothing more.

W

Workspace: your team’s home. You arrange work with boards and lists, and keep the labels, fields and Landscape.
Wrap-up: what you ask a coding agent for at the end of a session. It writes the notes the session should leave.