What Snapchat Will Not Recommend or Pay For.

Snapchat publishes two sets of rules that decide how far a brand’s content travels and whether it earns anything.

One governs recommendation, meaning whether Snapchat shows your content to people who have not already added you. The other governs monetisation, meaning whether that content is eligible for payment.

Most teams have read neither. They are short, they are specific, and several of the disqualifying behaviours are things a high-volume publishing operation does by default.

The version most people have heard is that Snapchat has banned AI content. That is not what the policies say, and the real position is more useful.

Recommendation and monetisation are two different gates

The Content Guidelines for Recommendation Eligibility cover content that travels beyond your friends and subscribers: the Stories tab, Spotlight and the Map. Snapchat’s framing is that such content must meet additional, stricter standards to be eligible for algorithmic recommendation.

The Creator Monetisation Policy is a separate document, and governs payment.

They are not parallel tracks. The monetisation policy requires creators to follow the Community Guidelines, the Terms of Service, the Snap Monetisation Terms, the Spotlight Terms and the recommendation guidelines. Recommendation eligibility is a precondition of getting paid.

That produces the failure brands rarely diagnose. Content can be perfectly allowed on the platform, quietly ineligible for recommendation, and therefore earning nothing, with nothing taken down and no notification sent. It is one more reason Snapchat monetisation is now an operations game rather than a content one.

What Snapchat’s AI rule actually says

The Creator Monetisation Policy treats AI as an editing tool, not a disqualifier. Content made using sophisticated editing or AI-based tools is monetisable, provided it clears the conditions the policy sets out:

• It is original, entertaining or informative.
• It is not misleading.
• The use of AI is disclosed anywhere in the content or the creator profile.

Nothing there bans AI assistance. A team using generative tools for a cutaway, a background plate or a voice clean-up sits inside the policy, as long as it says so somewhere.

Where the disclosure has to go

That phrase, anywhere in the content or the creator profile, is broader than most teams assume.

A profile-level disclosure counts. You do not have to label every post. One clear line on the public profile stating that AI tools are used in production satisfies the wording as written.

A per-post disclosure counts too, and is safer when AI use is uneven across your output, because a profile note implies the whole account.

What Snapchat does not supply is a required format, wording or placement. No approved phrase, no official tag. So make disclosure a standing decision rather than a judgement call at publish time: pick the level, write the line, apply it identically every time.

Disclosure does not rescue fully generated video

Here is the part the coverage flattened, and it changes what a disclosure is worth.

Under the Quality section of the recommendation guidelines, AI-generated content appears in the list of what is not eligible for recommendation. Snapchat’s position is that its ranking algorithm rewards human-made content over wholly AI-generated content created outside of Snapchat, even when that content carries transparency disclosures.

AI-generated content created within Snapchat is eligible for recommendation, and Snapchat attaches transparency indicators to it.

Snap set this out publicly on 31 July 2026: from that month, wholly AI-generated videos are no longer eligible for recommendation on Spotlight, while content enhanced or edited using Snapchat’s own AI creative tools continues to be eligible.

So two details decide where your content lands. Whether the video is wholly generated or merely AI-assisted, and whether the AI ran inside Snapchat or outside it. Disclosure is a condition of being paid. It is not a route back into recommendation for a fully generated video made elsewhere.

What Snapchat has not spelled out

Snap’s announcement was framed around Spotlight. The quality guideline it rests on sits inside the wider recommendation rules, which also cover the Stories tab and the Map, and no surface by surface breakdown exists.

There is also no stated threshold for where assisted ends and wholly generated begins. A live-action clip with a generated background and a generated voiceover is not obviously either.

Our reading: treat the whole recommendation estate as covered, and treat any asset with no filmed or photographed source material as wholly generated until Snapchat says otherwise. Being conservative costs you one format. Being wrong costs you distribution, with no appeal queue behind it.

Reposting has a specific test

If you post someone else’s content, the monetisation policy requires you to add to it in a valuable, highly transformative way. The recommendation guidelines say it from the other direction, ruling out content you did not create and have not transformed creatively, plus low-effort reaction content that is merely a pretext to repost someone else’s clip.

Snapchat is unusually concrete about what clears the bar:

• Reactions where your face is visible and you are audibly speaking, rather than nodding or pointing, and where the reaction is relevant to the source material.
• Reviews that use clips as context for your own assessment.
• Compilations whose editing shows real editorial judgement.
• Properly attributed social clips carrying your own commentary on a newsworthy event.

And what does not:

• Unaltered clips from television, film or music videos.
• Re-uploading someone else’s social post without transformation.
• Mechanically generated reactions, stock footage and template-based visual effects.
• Content assembled purely by automation or formula, with no creative effort behind it.

The format most exposed is the cheapest one to produce: found clip, added caption, published. Working out which formats clear Snapchat’s bar, before you build a series on top of one, is what platform optimisation work is for.

Repetition is judged across the account

Not monetisable, per the policy: re-posting the same or similar content over and over and presenting it as new. Re-using the same tile image repeatedly sits in the same list. The recommendation guidelines separately rule out repeatedly posting the same content, whether your own or someone else’s, with minimal creative differences.

Read the word similar carefully. This is not a duplicate-file test.

A weekly format built from one template with a swapped clip and a swapped headline is similar by design. So is a series where the tile changes colour and nothing else. Volume is not the problem. Low variance between items is.

Snapchat also says it evaluates accounts holistically, using a mix of human and algorithmic moderation. The unit of judgement is the account and its pattern, not the single post a producer is worried about.

The rules that catch busy teams by accident

Several disqualifiers have nothing to do with editorial quality and everything to do with production hygiene:

• Substandard video: blurry, low-resolution or overly pixelated imagery.
• Wrong orientation, where a viewer would have to turn their phone from vertical to horizontal.
• Video that mistakenly has no audio.
• Tile images unrelated to the content, including unrelated celebrity photos.
• Illegible text on a partner tile.
• Promoting your accounts on other social platforms.
• Off-platform links promoting external messaging services.
• Flashes or strobes without a photosensitivity warning.
• Shocking imagery used as engagement bait, unfounded rumours presented as fact, and long-past events presented as current.

The cross-promotion one catches people out. Pointing Snapchat viewers at your other channels is a normal growth instinct, and it is named in the guidelines as something that costs you recommendation eligibility.

Why this bites the process before it bites a post

None of these is really a post-level rule. Each one is a default set somewhere upstream.

The transformative test is decided when someone chooses a source, not when they write a caption. The similarity risk is built into a template, so it gets set once and repeated fifty times. The AI disclosure is a settings decision about an account, not an editorial one about a video.

A team shipping at volume cannot catch any of it in review. The queue looks at one post at a time, and every rule above is about the pattern across many.

The rules are account-level. Almost every approval process is post-level.

What good looks like

• Every recurring format has a named transformative element that is present by construction. The presenter is on camera and speaking, or the edit carries an argument.
• AI use has one declared position: which tools, at which stage, and where the disclosure sits.
• Nothing without filmed or photographed source material goes to Spotlight.
• One person owns the ratio and can say, without checking, how much of the last thirty days was genuinely new against near-duplicates and third-party clips.
• Eligibility is checked at format design and again on a monthly sweep, never post by post.

Put the rules in the workflow, not in someone’s memory

Five additions that make these rules survive staff turnover:

1. A source field on every asset: filmed, licensed or third party. Third party triggers the transformative check before the edit starts.
2. A standing AI disclosure decision per account, recorded wherever your brand guidelines live.
3. A tile check in the publish step covering legibility, relevance, and whether the image has run before.
4. A monthly similarity sweep across the last thirty days of output.
5. A named owner for platform rules whose job explicitly includes re-reading them each quarter, because these pages change without announcement.

Most of this belongs in the publishing checklist, not a separate policy document nobody opens. If your process has nowhere to put it, the social media approval process template is a reasonable starting shape.

How NBK thinks about platform eligibility

Platform rules are not a compliance chore. They are the terms on which distribution is granted, and among the cheapest performance levers a brand has: failing them forfeits reach you already paid for in production.

In our view, eligibility failures inside a busy content operation are almost always process failures, not judgement failures. Nobody decides to publish a near-duplicate. A template makes it inevitable. Across 46,000+ posts shipped, the pattern holds: the rules a team breaks are the ones nobody owns.

So that is the work. Turn published rules into defaults, fields and checks, so nobody has to remember them and no single person’s attention is the control.

Next step

If your team publishes steadily on Snapchat and the reach or payouts do not match the output, the constraint is usually upstream of the content. Audit your recurring formats against Snapchat’s published rules, format by format rather than post by post.

If that turns up more than you expected, NBK can help rebuild the workflow behind the content.