AI and agents
What can the BetterCMS agent do?
The BetterCMS agent proposes changes across many documents at once and hands each one back for review, field by field, with undo. A rename that touches forty entries is one run and one review rather than forty edits. The agent works inside the same roles, approval stages and audit log a person does.
The agent also drives the page generators: a set of SEO pages from a keyword list, or one account based page per target company. Both arrive as drafts and go through the same review as anything else.
A run is a change set, not a chat
The unit of work is a set of proposed changes across many documents, produced in one go and reviewed in one go. Renaming a product across forty entries is one run and one review rather than forty edits and forty chances to miss one.
That framing is the difference between an assistant and an operator. An assistant helps you write a paragraph. An operator changes the site and shows you what it changed.
Review is field by field
Every proposed change shows its before and after on the field it touches, per document. You can approve a whole document, approve a subset of its fields, or discard it. Undo is available afterwards as well.
The reason to review at field level rather than document level is that a forty document run is usually thirty eight good changes and two the agent misread, and a document level yes or no forces you to choose between shipping the two or redoing the thirty eight.
The generators
Two of them ride the same runs machinery. One builds a set of SEO pages from a list of keywords, pasted in or uploaded as a CSV. The other builds one account based page per target company.
Both produce drafts, and both go through the same review and the same audit log as anything else. A generator that published straight to the site would be the auto-publish this product does not do, wearing a different name.
Also in ai and agents
Still have a question left
Thirty minutes with the product team, and a straight answer.