When your AI tooling writes the brief, the outline or the script, that involvement can call for its own disclosure — separate from, and in addition to, the #ad. The platform's AI label doesn't cover it. Most teams are tracking one layer in a spreadsheet and the other nowhere at all.
Buy Now — $19One-time payment · works offline · nothing leaves your browser
The mistake is treating these as the same box. They have different triggers, and an engagement can be clean on one and exposed on the other.
Payment, gifted product, commission, an ambassador retainer, an employee posting about their own employer. Was the connection disclosed clearly and conspicuously in the content — not in a contract, not only in the bio, and not only through the platform's partnership tool?
Did the brand supply the prompt or outline? Was AI output used with only light editing? Did the creator independently verify or rewrite the claims? Do the claims go beyond what that creator would credibly say unprompted? Two or more risk factors and the app flags a separate AI-involvement disclosure as indicated.
One record per published piece, filled in before or at publication.
Creator, handle, platform, brand, campaign, publish date and content type, across paid partnerships, gifted product, affiliate, ambassador, employee posts, equity and contests — with the material-connection default set per engagement type and overridable per record.
Spoken, on-screen, in the caption above the fold, below the fold, bio only, or the platform's partnership tag. The weak placements are recorded as weak rather than counted as a clean disclosure, because that distinction is the whole argument.
Keep the AI draft against the as-published wording, or at minimum record where the draft lives. If nobody can see what the tool generated versus what the creator changed, there is no materiality argument left to make later.
Route anything ambiguous to legal with a timestamp and a note, and it stays at the top of the register until it's resolved. Print a dated pack: process statement, summary, the full register and a detail sheet per engagement.
Five engagements from one creator programme. The gaps sort themselves to the top.
| Creator | Platform | Published | Layer 1 · payment | Layer 2 · AI | Audit trail |
|---|---|---|---|---|---|
| Ravi ChandraPaid partnership · flagged to legal | TikTok | 23 Jul 2026 | Disclosed | Disclosure missing | Retained |
| Duncan ReyesAffiliate / commission | YouTube | 9 Jul 2026 | Not disclosed | Disclosed | Draft not retained |
| Mei TanakaGifted product | 16 Jul 2026 | Platform tag only | Below threshold | Retained | |
| Priya NandakumarPaid partnership | 24 Jun 2026 | Disclosed | No AI involvement | Not applicable | |
| Sable CreativeEmployee posting | 11 Jun 2026 | Disclosed | Below threshold | Retained |
Every piece of advice published on this says the same thing: build a process that tracks it before a creator hits record. None of them hands you the thing to track it in, so teams end up with a spreadsheet that has one disclosure column and no audit trail. This is that thing, for $19, once.
Buy it, open it, start logging. No account, no subscription, no seat count.
No. A platform label tells the audience the media was made or altered with AI. It says nothing about who shaped the claims or who paid for them, and the published guidance is explicit that sponsored AI-assisted content can need both disclosures, separately. The app keeps them apart for exactly that reason.
No, and it says so on every screen and in the printed pack. Two or more risk factors raise a flag so a human looks at it. An engagement under the threshold can still call for a disclosure, and one over it isn't automatically non-compliant. These are factors to weigh, not a verdict.
Three things a spreadsheet column doesn't give you: the AI question as its own assessed field rather than a note, disclosure placement graded so a bio-only or platform-tag-only disclosure isn't counted as clean, and the draft-versus-published record that makes a materiality argument possible months later.
Nowhere. It's a single HTML file that stores everything in your own browser. You'll be typing in creator names and draft scripts, so that matters. Export a JSON backup or a CSV whenever you want a copy or need to hand the register to someone else.
No. It's a record-keeping and checklist aid. It helps you write down what you did and shows you where the record is thin. What that means for your programme is a question for your own counsel.