
Website and CMS content updates
Small website changes wait weeks behind whoever holds the CMS login. An agent takes each request to a saved draft with the before and after side by side, and a named reviewer decides what goes live.
The short answer
You keep website content current by letting an agent take each change request to a saved draft in your CMS, with the before and after side by side.
It edits only drafts, checks links, alt text and redirects, and logs what changed. A named person reviews every draft and publishes it.
Why do small website changes sit for weeks?
The request is small: a new price, a team bio, an event date, a broken link someone noticed on a call. It lands in an inbox or a chat thread and waits for whoever has the CMS login and a free afternoon. By the time it ships, two more requests are queued behind it, and the wrong page has been wrong in public the whole time. Nobody owns the queue, so nobody can say what was asked for, what changed, or when.
Who is this for?
A business whose website runs on a CMS, where marketing or operations staff ask for changes and one or two people can make them. It suits a steady flow of small edits to pages that already exist. A redesign, a new template or a change to how pages are built is design and development work, and it stays with the people who do that work.
What changes when an agent takes the request?
| Part of the update | Request in an inbox | With an agent drafting |
|---|---|---|
| Where the request lives | An email or a chat message that scrolls away | A written request with the page and the change named |
| Who makes the edit | Whoever has the login, when they get to it | The agent, in a draft, as the request arrives |
| What the reviewer sees | The live page, after the fact | The current copy and the proposed copy together, with check results |
| What goes live | Whatever was typed into the editor | Only a draft a named reviewer published |
| The record | Scattered across threads, if it exists | One log entry per change: who asked, who approved, when |
| Undoing a mistake | Someone remembers what the page said before | The earlier version is restored from the page history |
How it works, step by step
Each step takes one request from arrival to a decision. The agent works inside the access you grant and nowhere else.
Take the request in writing
Find the page and copy what it says now
Write the change into a draft
Check what the edit touches
Human gate: the page owner publishes
Log the change
Keep the queue in view
What should an agent never do on your website?
Publish on its own
It saves drafts. Publishing is the reviewer's action.
Change design or code. It edits copy, links, images and page settings inside the layouts you already have. Themes, templates, plugins and code stay with your developers.
Write claims nobody approved
Prices, offers, testimonials and legal wording come from the requester or the page owner, word for word.
Hold more access than the work needs
Where your CMS has a role that can save drafts without publishing, the agent's account uses that role.
Make a change you cannot reverse. The page history keeps the earlier version, and the log names the request that changed it.
Where this fits
This is a Tier 1 engagement, Automate a Workflow: one process, end to end. Website updates make a sensible first candidate when requests arrive often, follow a pattern and already end with someone checking the page. Run the readiness checklist against your own request flow first. Start in shadow mode, with the agent preparing drafts beside the person who does the updates today, so you can compare its drafts with theirs before any reach a reviewer alone. The use case library has the neighbouring workflows.
How does this differ from a content pipeline?
| Question | Content production pipeline | Website content updates |
|---|---|---|
| What starts it | A content calendar and a brief | A request to change a page that already exists |
| What it produces | A new piece, from outline to publish-ready draft | An edit to existing copy, links or images |
| What it checks | Brief, voice and schedule | Links, alt text, headings and redirects on the changed page |
| Who signs off | The editor who owns the calendar | The person who owns the page |
Related reading
- Content production pipelines that ship on schedule: new pieces taken from brief to a publish-ready draft.
- FAQ and help content that updates itself: keeping help articles in step with the questions customers ask.
- Approval gates: why each sign-off belongs to one named person.
