Brand and Style Check Agent · Use case
How can we use an AI agent to check a draft against our brand and style guide on our own infrastructure?
A brand and style check agent reads one draft against your brand and style guide and lists each departure: a product name written the wrong way, a term the guide retires, a tone the guide rules out, a heading in the wrong case. Beside each it quotes the line and the rule. The editor accepts or overrides each point. The agent changes nothing in the draft, and it does not call a draft on-brand: it reports what it found against the guide you indexed, and an empty list is not an approval.
The guide runs to sixty pages.
The draft is due today.
- ✗ 01A brand and style guide is long, changes often, and is read closely by few of the writers, agencies and colleagues who draft under it.
- ✗ 02An editor reads every piece for the same points, and the judgement that matters, when to break a rule, stays with the editor.
Where it stops for a person
Always. The editor accepts or overrides each point.Always. It lists, it does not correct and it does not pass a draft: no result reads "on-brand" or "approved". Where the guide is silent it says so and offers no rule of its own. The check is of the draft, not of its writer: it keeps no tally of who departs from the guide, and gives no score for a writer or an agency. Facts and claims are not its subject: those go to the claim check.
It lists each departure with the rule.
The editor accepts or overrides.
Step · 01
The brand and style guide is indexed as a knowledge source, with its current version marked.
Step · 02
A writer or an editor uploads the draft.
Step · 03
The agent reads the draft against the guide's rules on names, terms, spelling, tone and layout.
Step · 04
It lists each departure with the line, the rule and where the rule is in the guide, and marks the points where the guide is unclear or silent.
Step · 05
The editor accepts or overrides each point and corrects the draft.
Reads from · connectors
The one draft uploaded; and your brand and style guide, glossary and list of product names, indexed as a knowledge source with each reader's permissions kept.
Acts through · MCP servers
Nothing in your systems. It writes a list. It has no tool that edits the draft or publishes anything.
What you set up
- Index the current brand and style guide, the glossary and the list of product names as a knowledge source, and remove the versions no longer in force.
- Say which rules are firm and which are the editor's to waive.
- Build the agent in Katonic Studio and test it on invented drafts, and on published pieces your editors have already marked up.
- Publish it to one content team first.
Where it runs
The agent, its knowledge store and the model run on infrastructure you own or control. You choose the model and serve it yourself, so questions, documents and answers are not sent to a public AI service. The draft, the guide and the list stay inside the company.
How you get it today
This page is labelled Use case: the agent is built in Katonic Studio. You describe the job, connect what is named above, add test questions and release the agent once it passes them. When the Catalog holds a tested template for this job, this page will carry the Template label and show how to take it.
Does it correct the draft?
No. It lists each departure with the rule. The editor accepts or overrides each point and corrects the draft.
If the list is empty, is the draft on-brand?
Not for certain. An empty list means the agent found no departure from the guide you indexed. The editor decides.
Is the list used to rate writers?
It is not built for that. It is a list of lines and rules, with no count or score by person.