Short answer
Quality checks are automatic tests that run on every piece of content before a person is asked to approve it. In AI Marketing Master they look for prices or offers you have not approved, wrong business names or phone numbers, broken links, missing titles and images, banned phrases and copied text. A failed check blocks approval, so reviewers spend their time on judgement, not on proofreading.
Why a reviewer alone is not enough
When an owner approves twenty posts on a phone between appointments, small errors get through: an old phone number, a discount that ended last month, a link to a page that was deleted. These are exactly the mistakes AI tends to make, because it writes confident text that looks right.
Machines are good at boring, exact checks. People are good at judgement. Splitting the work this way makes each approval faster and safer.
What gets checked
- Claims: any price, discount, refund or warranty promise, delivery time, result or certification must match a claim you have approved. If not, it is blocked and listed as a missing claim for you to approve or remove.
- Business identity: the name, phone, address and website in the text must match your Business Profile exactly. Look-alike web addresses and typos are flagged.
- Links: every link must be valid and must actually open. A link that cannot be checked is marked as unable to check, never as a pass.
- Required fields and lengths: a blog post needs a title, meta title, meta description and web address, and meta text must fit sensible limits.
- Media: blog posts and pages need an image, and each platform's file type, size and video length rules are applied.
- Banned phrases and topics: words you never want used always fail. Topics you want to avoid are flagged.
- Sensitive wording: medical, legal and money language is sent to a person for review instead of passing quietly.
- Copied text: content too close to your earlier posts is flagged, after ignoring your business details and calls to action.
Three quality levels
Each business chooses Standard, High or Strict. Higher levels are less tolerant of text that repeats earlier content. On Strict, every item also needs a person to review it, even if every automatic check passes. Clinics, financial advisers and agencies with demanding clients often choose Strict; a café posting daily specials may be happy with Standard.
What happens when a check fails
The item shows a red Quality failed status with the reason written in plain words, for example that the phone number does not match your profile. Approval is not possible until it is fixed. You can edit the text yourself or ask Milo to rewrite the part that failed, and the checks run again on the new version.
Checks also run a second time just before publishing. If your approved facts changed in between, for example an offer expired, the content is stopped before it goes out.
What the checks cannot judge
Automatic checks cannot tell whether a post is interesting, whether the advice is wise for your customers or whether the tone fits your brand. They also cannot confirm that a statement which is not a claim, such as a description of a process, is accurate. That is still the reviewer's job, and it is why approval stays with a person.
What it does not do
- Checks catch factual and format problems, not weak ideas or the wrong tone.
- A link that cannot be reached at the time of checking is shown as unable to check and needs a person to look.
- Checks are only as good as your Business Profile: keep approved facts current.
Questions people ask
Can I skip a failed check if I am sure the content is right?
No. A failed mandatory check blocks approval. Fix the cause instead, for example by approving the claim in your Business Profile if it is true.
Do the checks work on social posts too?
Yes. They run on blog posts, website pages, social posts and Google Business Profile posts, with rules suited to each type and platform.
Are the checks done by AI?
The core checks are fixed rules, so the same text always gets the same result. That makes them predictable and easy to explain.
Do checks slow publishing down?
They take moments. Most of the time saved comes from reviewers no longer needing to hunt for small factual errors.
We only describe what the product really does today. Results depend on your business, your market and many things no tool controls.