Legal review on social media exists to answer one question: can we say this? Not whether the post is on-brand or well-written, but whether the claim is substantiated, the comparison is defensible, and the post will not earn a regulator letter or a competitor complaint. The failure mode most teams fall into is binary: either nothing touches legal and risky claims slip out, or everything does and the queue grinds to a halt.
This guide is about the claims layer specifically, not general approval. The aim is a workflow where 90% of posts never need a lawyer, and the 10% that do arrive at legal already evidenced and easy to clear.
What actually triggers legal review
Most social posts carry no legal risk. The job is to reliably catch the ones that do. Build a trigger list so authors self-identify risky posts instead of guessing:
- Performance or superiority claims — “the fastest,” “number one,” “saves you 40%”
- Comparisons that name or clearly imply a competitor
- Testimonials and reviews used as endorsements
- Health, financial, legal or safety claims in regulated categories
- Pricing, guarantees and “free” offers with conditions
- Anything implying a result a customer should expect
A post hitting any trigger goes to the legal lane. A post hitting none follows pre-approved rules and self-publishes. This is the same risk-tiering that powers any social media approval workflow , applied to legal exposure rather than brand voice.
Substantiation is the whole job
A claim is only as safe as the evidence behind it, and legal cannot evaluate what they cannot see. Require a substantiation note attached to every triggered post: the claim, the proof, and the date the proof was valid.
| Claim in the post | Required substantiation |
|---|---|
| “Cut onboarding time by 40%” | The study or customer data, sample size, time period |
| “The fastest tool in the category” | Benchmark methodology, what was tested, when |
| “Loved by 10,000 teams” | Source of the count, as-of date |
| “Rated 4.8 stars” | Platform, review count, capture date |
A claim with no attached evidence is not “pending review,” it is rejected by default. This one rule eliminates the most dangerous category: confident posts nobody can back up. It also speeds legal up, because counsel reviews the evidence rather than chasing the author for it.
Pre-clear language so legal scales
The way to keep legal from becoming a bottleneck is to do the work once. Maintain an approved-claims library: phrasings legal has already cleared, with the conditions under which each is safe to reuse. “Trusted by leading teams” might be pre-cleared unconditionally; “saves 40%” pre-cleared only with the linked study and only until the data is twelve months old. Authors pull from the library and skip the queue entirely. Legal only sees genuinely new claims.
Pair the library with standing disclaimer rules: which offers need “terms apply,” when a testimonial needs “individual results vary,” how to label paid partnerships. Encoding these as rules means the team applies them without a lawyer in the loop every time.
Route by risk, not by reflex
Three lanes keep the queue honest:
- Self-serve — no triggers, follows pre-approved rules and the claims library. No legal touch.
- Light review — a trigger fires but the claim matches a library entry with conditions met. A marketing lead confirms the conditions; legal is not pulled in.
- Full legal review — a new claim, a comparison, or a regulated-category statement. Counsel reviews the substantiation and approves, edits the wording, or rejects.
Most posts land in lane one. If your lane-three volume is high, the claims library is too thin, not your lawyers too slow.
A worked example
A SaaS team wants to post: “Switch to us and cut your reporting time in half. See how [Customer] did it.” The comparison-and-result claim trips two triggers. The author attaches the case study, the 52% measured reduction, and the date. Legal edits “in half” to “by more than half” to match the data exactly, approves the customer quote because there is a signed release on file, and adds the phrasing to the approved-claims library scoped to that case study. Next quarter, a similar post reuses the cleared language and never touches the legal queue.
Metrics that matter here
Track legal queue time (trigger to decision), block rate (share of reviewed posts rejected or rewritten), and library reuse rate (share of claims served from pre-cleared language). Rising reuse and falling queue time mean the system is learning. A high block rate is a signal to fix briefs upstream, because authors are repeatedly proposing claims that cannot be backed.
How Utin fits
Utin is being built to scan your website for the claims you already make and carry them into social drafts with the source attached, so substantiation is present from the first draft. Posts that hit a trigger route to the legal lane automatically; cleared phrasings feed an approved-claims library the next draft can reuse. The intent is that legal sees fewer, better-evidenced posts and the rest of the calendar never waits on them. You can register interest in an early pilot.
For the documentation and record-keeping side of regulated work, read social media compliance workflow . For org-level ownership of who can approve what, see social media governance . When AI generates the drafts, keep AI posts with human review close, since unverified AI claims are exactly what this lane exists to catch.