The short version
The documented checks cover:- Promotional claims: codes, discounts, prices, and urgency language that need a real offer behind them.
- Template placeholders: unfinished stand-ins for a product, business, location, or source.
- Incomplete output: empty or truncated drafts.
- Writing patterns: repetitive openings, filler transitions, dash use, and keyword repetition flagged by the internal scorer.
- Image safety: moderation of generated images.
How promotion checks support review
A customer may act on a price or discount in an article. Supply the actual offer, eligibility, and dates in your business information, then check that the published wording matches those details.Promo and coupon codes
The documented workflow treats suspected fabricated codes, such as “SAVE20” or “WELCOME10”, as a reason to reject and rewrite a draft. The fallback can use general wording such as “contact us for current offers” rather than an unsupported code. That describes the intended check, not a guarantee that every invented or expired code will be detected. If your business has a genuine offer, verify its exact wording and conditions rather than assuming a familiar-looking code is valid.Specific discounts, prices, and urgency
The documentation describes flagging figures such as a discount percentage or monthly price, along with deadlines such as “ends Friday”. A flag asks for verification; it does not establish that the figure is false. Equally, an unflagged figure is not proof that it is current.Evidence to review for a promotional claim
| Claim in the draft | Check before publishing |
|---|---|
| Price or fee | Match the current product, currency, billing basis, and included scope. |
| Discount or code | Confirm the real offer, eligibility, redemption method, and any exclusions. |
| Sale deadline | Confirm the actual date and relevant timezone; remove invented urgency. |
| Customer quotation | Use supplied, approved wording and confirm permission and context. |
Leftover template placeholders
The placeholder check looks for unfinished text such as[your product], [competitor], or [insert relevant statistic]. The documented response is to rewrite using the business context. Read the replacement: a concrete-looking sentence still needs a real fact behind it.
The documentation distinguishes placeholders from build markers and references:
- Internal-link markers, such as
[INTERNAL:...], are handled by the linking workflow. - Image markers, such as
IMAGE_1, belong to image placement. - Numbered references, such as
[1], can be legitimate citations.
Integrity rules for generated content
The writing instructions prohibit fabrication of:- Promo codes, discounts, sale dates, prices, and contact details.
- Statistics, research, or named studies without supporting sources.
- Testimonials, reviews, customer quotations, or implied customer experiences.
- Awards, certifications, rankings, partnerships, or endorsements without evidence.
User Direction does not authorize fabrication
User Direction steers the topic, audience, and emphasis of an individual post. The documented instruction hierarchy keeps content-integrity rules above that direction. Asking for a made-up promotion or testimonial does not make it appropriate to publish. The generation documentation describes using general promotion language when a specific offer is not supported by the business context. Review the output rather than assuming the instruction hierarchy is an infallible filter on every publishing path.Minimum-quality gates
Two documented checks help identify incomplete or weakly supported output:- Minimum content length: the documentation describes a floor of about 500 characters, with shorter output treated as truncated and sent for rewriting. This is an internal completeness check, not a Google word-count requirement or proof of useful coverage.
- External links: the rubric recommends two or more external sources. Fewer links can trigger a warning or reduce the link score, without necessarily blocking the draft. Read the sources and check what they support; link quantity is not factual verification.
Writing-pattern penalties
The internal scorer can deduct points for stock openers, repeated filler transitions, heavy dash use, repeated paragraph openings, or excessive keyword repetition. These are prompts for editorial review. A quotation, technical term, or deliberate repetition can be appropriate in context. The documentation also describes trimming excess dashes during writing. Check whether an edit improves meaning and flow rather than treating punctuation as evidence of factual accuracy or a search-engine penalty. Google’s spam policy addresses manipulative keyword repetition; it does not make this internal writing rubric a Google assessment. Pattern deductions are described as score changes rather than hard publication blocks. Inspect the relevant passage before deciding what to revise.Image safety
Generated images are described as passing through moderation before attachment. A blocked image can produce a message asking you to try a different prompt. Check the post’s current status and whether text or images need attention; the generation guide distinguishes image-only retries from rewriting the whole article. Moderation does not verify a depicted product, customer, professional credential, or procedure. Review an image’s relevance and accuracy, along with its alt text, before publishing. Do not use an illustrative image as proof of a real customer outcome.Review before relying on automation
Before publication, confirm:- Offers and contact details match current approved business information.
- Important factual claims have sources that support the statement.
- Customer quotations and proof have permission and are represented accurately.
- Placeholders, links, citations, and images work in the rendered draft.
- The selected publishing mode and review responsibilities fit the content’s risks.
