Traceable Agents: AI Agents as WordPress Users
We are working on one of the biggest updates we have ever built: AI Agents as WordPress Users.

This idea has been around for a long time. We actually started working on it quite a while ago, but we decided to put it on hold so we could focus our resources on improving the user experience. A few days ago, we rebuilt a the pull request, redesigned several parts of the feature, and now it is finally in the testing phase.
And honestly, we believe what is coming is going to be huge.
AI Needs an Identity
Most WordPress websites are managed by a single administrator. One person writes the content, manages plugins, updates the site, and takes care of everything.
Of course, there are also large websites, such as online newspapers or enterprise platforms. where many people work together every day.
No matter the size of the website, the idea behind this feature is the same.
Today, there are many AI tools available for WordPress. They create content, edit posts, delete data, update settings, and automate tasks. But they all share the same problem: they are not traceable.
Even worse, they usually don’t have a proper permission system. If an AI agent has access to the wrong tools, a simple mistake could cause serious damage to your website.
AI Agents as WordPress Users
Our next release focuses on three key concepts:
- Traceability
- Permissions
- Identity
Instead of running as anonymous AI assistants, agents will become actual WordPress users.
You will be able to create a dedicated WordPress account for each AI agent, define its identity, configure its instructions, and decide exactly which tools and permissions it should have.
Each agent will have a clear role, just like any other member of your team.
Every Action Is Recorded
Because agents work through real WordPress users, every action becomes traceable.
You’ll know:
- When an agent edited a post.
- What content it changed.
- What it deleted or created.
- Which actions it performed during its work.
This gives you complete visibility into everything your AI agents do, making automation much safer and easier to trust.
Imagine a Different Workflow
Picture this.
You’re in your WordPress Desktop (yes, the place we all used to call wp-admin).
There is a new section called Agents.

Inside, you have several specialized agents:
- An SEO expert.
- A grammar and style reviewer.
- A translator.
- An image optimization assistant.
- Or any custom agent you create.
Now imagine you have a post ready to review.
Instead of opening prompts or copying and pasting text into different tools, you simply drag the post onto one of your agents.

The agent immediately starts working.

Or imagine an even more natural workflow: you simply right-click on a post and select “Send to Agent Spanish Translator”. The chat window will open, the agent will start to work autonomously, and it will create a Spanish version of the post, exactly with everything you have asked for defined in its behaviour.

Once it’s finished, every change is recorded. You can review exactly what it modified, what decisions it made, and when the work was completed.
Everything is transparent.
Everything is traceable.

This Is Just the Beginning
This update is about much more than letting AI edit WordPress content.
It’s about giving AI agents a real identity inside WordPress, with their own permissions, responsibilities, and complete activity history.
We believe this will unlock entirely new workflows that simply weren’t possible before.
And this is only the beginning.

















