Content Rules

AI content rules that actually enforce.

Persistent, plain-English guardrails that apply before the AI ever writes — not as after-the-fact filters. Keep AI output on-brand, on-message, and brand-safe across every platform you post to.

Pre-generation enforcement · editable any time
velocipost.com/dashboard/settings/rules
Content rules
6 active
Always mention 24/7 emergency availability on service posts
Never mention competitor names or services
Educational tone — never salesy
Avoid pricing claims or discount language
Include CTA every 3rd post: "DM us to schedule"
Saved · applied to every generation
Why rules matter

AI without guardrails is a risk.

Generic AI writes whatever it thinks sounds good. Most of the time, that's fine. Occasionally, it's a post about a competitor, an off-brand tone, a pricing claim you can't back up, or a disclaimer that never made it into the draft. Small businesses have embarrassed themselves online for less.

Content rules close that gap — not by reviewing what the AI wrote after the fact, but by telling it what it can and can't write before a single word appears. Rules are the difference between "we use AI for social media" and "our AI sounds exactly like our brand."

Competitor mentions

Generic AI doesn't know not to name-drop your rivals. One pass through and your account is recommending the competition.

Unbacked claims

"Best in the area," "guaranteed results," "#1 rated" — AI reaches for hype language automatically. Your industry may prohibit it.

Off-brand tone drift

Your brand is warm and educational. The AI writes an aggressive sales pitch. It happens. Without tone rules, it keeps happening.

Missing disclaimers

Regulated industries require specific language on every public communication. AI forgets unless you make the rule explicit.

The three rule types

Three kinds of rules, one system.

Every rule you write falls into one of three categories. Combine them to build the guardrails that fit your brand.

Must-include

Phrases, CTAs, disclaimers, or topics that have to appear. The AI treats them as non-negotiable elements that belong in the output.

Common examples
"Always mention 24/7 availability on service-related posts."
"Include the CTA 'Book a free consultation' on every 3rd post."
"Always use the hashtag #YourBrand on Instagram posts."

Must-avoid

Topics, terms, competitor names, or specific claims that the AI can't write. Enforced absolutely — if it violates a rule, the output gets rewritten.

Common examples
"Never mention competitors by name or service."
"Avoid superlatives like 'best,' 'top,' or '#1.'"
"No specific pricing claims or discount language."

Tone directives

Style guidance that shapes the voice of every generation. Not rigid rules — directional guidance that the AI respects across every platform and every post type.

Common examples
"Educational, not salesy — teach rather than sell."
"Professional but warm — not stiff, not overly casual."
"Use concrete specifics, not generic claims."
How rules enforce

Applied before the AI writes. Not after.

There's a world of difference between "filter the output" and "tell the model what it can write in the first place." velociPost does the second.

How other tools do it

Filters after the fact.

The AI writes whatever it writes. Then the tool scans output for problem words, flags what it catches, and hopes you'll notice the rest in review.

1
Model writes — no constraints
2
Filter scans output for banned terms
limited
3
Human review catches the rest — maybe
inconsistent
Output reflects the model's defaults — not your brand.
How velociPost does it

Rules first. Then the model writes.

Your rules are bundled into every prompt as explicit constraints. The AI reads them before it writes, so the output already respects them by the time it reaches you.

1
Rules load from your workspace
pre-load
2
Prompt sent with rules as constraints
bundled
3
Model writes within the guardrails
enforced
4
You review the pre-filtered output
final check
Output reflects your rules from the first draft.
Rules in practice

What a rule set actually looks like.

A handful of plain-English rules is usually all it takes. Here are two example configurations — no industry-specific templates required, just clear constraints that make AI output feel like yours.

5 rules · takes ~3 min to set up

A minimal rule set.

Include a specific call-to-action on every 3rd or 4th post.
Never reference competitors by name.
Avoid superlatives like "best," "top," or "#1."
Educational in tone — teach, don't pitch.
Use concrete specifics — never generic filler.
8 rules · more conservative brand

A tighter rule set.

Always include required disclaimer on service-related posts.
Use full business name once per post — no abbreviations.
No guarantees, promises of outcomes, or unverifiable claims.
Avoid urgency language ("act now," "limited time," etc.).
No specific pricing, discounts, or promotional language.
Never reference individual customer names or sensitive details.
Professional tone — never casual or colloquial.
Emphasize expertise and process — not outcomes or results.
Where rules apply

One rule. Enforced everywhere.

Content rules govern every AI generation in your workspace — not just posts. Set the rule once and it applies across every surface the AI touches.

Posts

Every scheduled and on-demand post, across all 11 platforms.

Replies

AI-drafted responses to comments in the engagement inbox.

Images

AI image prompts respect visual content rules too.

Video hooks

Word-by-word video text overlays follow the same rules.

FAQ

The usual questions.

What are AI content rules?

Content rules are persistent constraints the AI reads before every generation. They tell the model what must always appear (disclaimers, CTAs, specific phrases), what must never appear (competitor names, banned topics, specific claims), and how the tone should sound (formal, casual, educational, never salesy). They're rules, not suggestions — the AI respects them on every post and reply.

How do content rules ensure brand safety for AI content?

The rules apply pre-generation, not as after-the-fact filters. When it's time to write a post, velociPost bundles your rules with the request and sends them as explicit constraints to the AI. The model reads them as part of the prompt — so the output already respects them by the time you see the draft. You also review every draft before it publishes, which catches anything the rules didn't.

What kinds of rules can I set?

Three types: must-include rules (phrases, CTAs, disclaimers that have to appear), must-avoid rules (topics, competitor names, specific claims that can't appear), and tone directives (style guidance like "educational not salesy" or "professional but warm"). You can add, edit, reorder, or remove rules any time from Settings.

Do content rules apply to replies and comments, or just posts?

Both. Content rules govern every AI generation in your workspace — posts, AI-drafted replies in the engagement inbox, and image and video prompts. One rule, enforced everywhere the AI writes on your behalf.

Can I have different rules for different platforms?

Rules apply across all 11 platforms by default. If you need platform-specific constraints, you can add rules scoped to a specific platform — useful when LinkedIn needs a more formal tone than Instagram, or Google Business Profile posts need a specific disclaimer that Instagram doesn't.

What happens when the AI writes something that violates a rule?

Because rules are enforced pre-generation, outright violations are rare. If you do spot an output that drifts from a rule, reject it and add a short reason — that signal gets baked into future generations. If a specific rule is being violated repeatedly, tighten the wording of the rule and save. Changes apply to the next generation immediately.

How do I add or edit content rules?

Open Settings → Content rules. Add a new rule, choose its type (must-include, must-avoid, tone), write the rule in plain English, and save. Rules are editable, reorderable, and can be toggled on or off without deletion. Every change applies to the next generation.

Can I use content rules for compliance purposes?

Yes. Content rules are a useful layer for compliance-sensitive businesses — required disclaimers, banned claims, regulated terminology. That said, velociPost is not a compliance tool, and you should always review AI-generated content before publishing if your industry has legal or regulatory requirements. Rules reduce risk; human review closes the loop.

AI that stays on-brand.

Join the waitlist and we'll email you the moment early access opens. No credit card, no commitment.

We'll email you the moment early access opens.