About this case study
This models the setup step, not a quality guarantee. Better context produces drafts that are closer to usable; it does not remove the need for human editing and fact-checking.
Generic output is usually a briefing problem
A team tries the draft feature, gets back something that could describe any product in the category, and concludes the tool writes badly. Often the more accurate conclusion is that nothing in the workspace told it what the product does, who it is for, or what the brand refuses to say.
A human copywriter given the same brief would produce the same beige paragraph. The difference is that nobody expects a new freelancer to guess the positioning on day one, while software is expected to infer it.
The setup step exists to close that gap once, rather than re-explaining the brand inside every prompt. It is the highest-leverage half hour in the whole workflow, and it is the step most teams skip.
What to actually write in each field
Three fields carry most of the weight when a campaign is created:
- Start from the website URL so the workspace has a first pass at what the product does, then correct whatever it got wrong rather than accepting it.
- Write the product context as if briefing a new hire: what it does, who it is for, what it is not, and the constraint a stranger would get wrong.
- Describe tone by behaviour rather than adjectives — "answers the question in the first line" is usable guidance, "friendly but professional" is not.
- List the restrictions explicitly: no pricing claims, no competitor callouts, no invented statistics, no pushy calls to action.
- Generate three drafts immediately, note what still reads wrong, and fix the context rather than editing the three drafts in isolation.
Restrictions do more work than instructions
Positive instructions shape a draft loosely; restrictions prevent the specific failures that make a brand look careless. A pricing claim that is six months out of date, a competitor mention nobody approved, or a statistic with no source will each cost more to clean up than the draft saved.
Restrictions are also the part worth revisiting. Every time a draft has to be edited for the same reason twice, that reason belongs in the restrictions field instead of in someone’s memory.
Teams running several brands should treat this as a per-workspace decision. A workspace per brand keeps context, tone, and restrictions from bleeding across accounts, which matters most when the brands are superficially similar.
The modeled operational outcome
A half-hour of written context turns every later draft from a blank-page generation into a brief-driven one. The gain is fewer rewrites and a shorter edit; the writing judgment stays with the team.
The guardrail worth keeping
Good context makes wrong claims sound more convincing, not less. Specific numbers, feature promises, and comparisons still need verification before publishing, and the confident tone of a well-briefed draft is exactly what makes that easy to forget.
Frequently asked questions
Why do generated posts sound generic?
Usually because the workspace has no product context, no tone guidance, and no restrictions. With an empty brief the safest output is the most average one.
How specific should the tone guidance be?
Specific enough to be checkable. Describe sentence length, opening style, and whether the brand uses humour or jargon, instead of listing adjectives that every brand would claim.
Should each brand have its own workspace?
Yes, when the brands differ in positioning or tone. Separate workspaces keep context and restrictions from leaking between accounts, which is the failure mode agencies hit first.
Does better context remove the need to edit drafts?
No. It shortens the edit and reduces full rewrites. Verification of claims, original examples, and the final call on publishing all stay with a person.