All terms

THE INKORE GLOSSARY

AI agent

An AI agent is a system that prepares or carries out steps towards a given goal, using tools and within defined limits.

Also known as: agent · autonomous agent

Where a simple assistant answers a question, an agent chains steps together: reading a source, calculating, drafting, proposing an action. What defines it is its autonomy over those steps — and therefore the limits it is given.

Two things must be decided in advance: which tools it can use, and which actions require approval. Without that second rule, an agent can confuse preparing with acting.

IN A REAL SITUATION

For example.

An agent can prepare a progress update from chosen documents, then offer to send it to the team. Sending remains subject to a separate human approval.

IN INKORE

How Inkore applies it.

Inkore acts as an agent for internal steps: reading the chosen sources, preparing a document, breaking down a mission. For external actions, it stops and presents the action in “Review”.

See the product

The point to keep in mind.

Define its tools, its sources and the actions that require approval before you give it a goal.

Not to be confused with: AI companion

An AI companion is one way of designing an agent: it works beside you, prepares, and leaves the decision to you.

FREQUENTLY ASKED QUESTIONS

Can an AI agent act on its own?

Technically yes, if it has the permissions. That is why you need to decide in advance which actions require approval: sending, publishing, paying or uploading to a third party should always require it. In Inkore, external actions wait in “Review” until you confirm them.

Back to the glossary

ONE INTENTION IS ENOUGH TO BEGIN

Shall we make
a little room?

For your next idea. For what really matters. Inkore is in preview: access is requested by email.