Comments and versions
How commenting works on objects, documents and approvals, how to call an agent with @, and how versions and suggestions keep every change visible and reversible.
Almost everything in a workspace can be discussed where it lives. A comment is a message on the thing itself. An agent that you call in a comment answers there, and it never changes anything silently: it proposes a new version, a marked change or an approval.
What can I comment on?#
- Objects with a detail page: goals, initiatives, risks, decisions, sessions, tasks and the like. Each has one thread, called Conversation.
- Documents and deliverables: on the whole text or on a passage you select. A comment on a passage stays attached to it.
- Approvals: the advisor side can discuss an approval and ask an agent to revise it.
Write the comment, press Reply on an existing one to answer in the thread. On the Mac and iPhone the object field says: comment, @ for an agent or a person. You can edit or delete your own comment, with undo. To close a thread press Resolve; it can be reopened.
Who sees my comment?#
Comments on objects and on documents are shared by default: the client side sees them as far as it sees the thing. The advisor can tick internal only; such comments are marked internal and only the advisor side sees them. A reply takes over the visibility of the comment it answers. The advisor's comments on a draft are internal. Comments on approvals are always internal. A person with Read access cannot comment. People you mention and the people involved are notified; internal comments notify only the advisor side.
How do I call an agent?#
Type @ and the name, for example the Editor on a deliverable, or answer an agent's comment. The thread shows that the agent is working, then its answer appears in the same thread, drawn dashed to show it comes from an agent. Which agent reacts depends on what you wrote and on the object. When the advisor replies to an agent or comments on its proposal, it always reacts.
What does an agent do with my comment?#
It does not edit behind your back.
- On a deliverable or document: it writes a suggestion, changes marked in the text, which only the advisor sees and decides (Accept or Reject per change). When all parts are decided, one new version is created. The comments in accepted parts are answered with the version that includes them.
- On an approval: it writes a new version of the proposal and says in the thread what changed.
- On an object: it creates an approval for the advisor.
If the passage you commented on has changed, the comment stays with its old quote. If you or someone else edits a text while a suggestion is open, editing waits until the suggestion is decided. A new suggestion replaces an open one.
Versions: who changed what, when and why#
Deliverables and approvals keep every version with who made it, when and why. A change by a person is drawn solid; a change by an agent is dashed. Two versions can be compared. Restoring a deliverable version creates a new version on top; see Deliverables. Objects have a list of Changes, field by field, with changes that came from an approval marked. The advisor can undo the newest change; it then shows as undone.
Common questions#
Can an agent change my text without asking?
No. It proposes. Nothing changes until a person accepts.
Can the client see internal comments?
No. Internal comments are visible to the advisor side only.
Can I undo an agent's change?
If it is accepted in a deliverable, restore the earlier version. For an object, the advisor can undo the newest change in the history.
Why can the client not comment on a document?
A client can. Only a person with the role Read cannot comment, and she does not see documents anyway.