Three deterministic tools, one rule book. The ad audit runs 62 versioned patterns against your copy in the selected vertical; the landing-page audit checks the page your ad sends traffic to; the competitor benchmark scores you against up to 5 rivals with the same engine. Here's exactly what each one checks.
Each family maps to published ad-platform or regulator guidance. Every finding names the policy, quotes the full offending sentence, and ships with fix guidance.
Catches best, #1, top-rated, finest, cheapest, lowest prices, nobody beats our prices, unbeatable — with smart exclusions so "we do our best" and "best practices" don't get flagged.
Catches money-back, 100% guaranteed, iron-clad guarantee, that's a promise, on the house, satisfaction guaranteed — promises that need visible terms or they read as exaggerated guarantees.
Catches save you up to $X, save $500, money back in your pocket, pays for itself, cut your bills by X% — projected-savings claims that need a disclosed basis.
Catches are you struggling…, tired of YOUR old AC, you're overpaying, if you're like most homeowners, homeowners like you, your family's safety — Meta bans asserting or implying the viewer's personal attributes.
Catches 100% safe, completely safe, harmless to children and pets, chemical-free, all-natural, non-toxic, zero risk, kills … on contact, wipes out … on contact — absolute safety/efficacy claims regulators prohibit.
Catches ALL-CAPS shouting, excessive punctuation (!!), gimmicky symbols and emoji, repeated words, and phone numbers stuffed into ad text — Google-only, ad-copy only.
Catches inflated "free" offers, missing disclosures on conditions, and strikethrough "regular price" tricks — US region only.
Catches VAT-excluded prices and vague "from £X" pricing — UK region only.
Catches "free" offers with vague conditions or hidden extra charges — UK region only.
Catches unsourced testimonials and star ratings with no review platform named — UK region only.
Catches familial-status, disability, religion, and gendered language in housing ads — "no kids", "perfect for singles", "church-goers welcome". Real-estate vertical only.
Catches guaranteed outcomes, "painless" claims, unjustified expectations, and discount ads with no total fee disclosed. Dental vertical only.
Catches guaranteed case outcomes, uncertified "specialist"/"expert" claims, and misleading testimonials. Legal vertical only.
Catches false "sold" status claims and unverifiable years-experience claims in UK property ads — UK region only.
Catches "only 3 left", "offer ends tonight", "limited time", "act now" pressure tactics with no genuine basis.
Catches bare "insured" / "bonded" with no verifiable credential — the report says "couldn't confirm — verify manually" instead of quoting it as proof.
No mystery math. The score is arithmetic on the findings — and the verdict is one of three labels. The tool pre-screens; it never blesses an ad as safe to launch.
Enter the page link — we open it and run every check on what we find. If a page can't be opened automatically, you can paste its content instead. Six deterministic check groups, same HIGH RISK / NEEDS REVIEW / CLEAR verdicts, same full-sentence evidence. Run a page audit →
Finds every US-format number plus all tel: links. The link wins on conflict — a stale visible number next to a current tap-to-call link gets flagged HIGH.
State patterns (OR CCB, CA CSLB, AZ ROC, WA, TX, FL…) plus generic numbered-license formats. A bare "licensed" with no credential is never quoted as proof — the report says "couldn't confirm — verify manually."
Compares the page title/headline name against body mentions (case- and punctuation-insensitive). Mismatches confuse visitors and hurt local search.
Counts call, quote, booking, contact, and form CTAs. Zero CTAs is a HIGH finding — visitors are never told what to do next.
Years in business, reviews and ratings, guarantees and warranties, insured/bonded wording — recorded as strengths when found, never invented.
Looks for "serving" areas, street addresses, and ZIP codes. A page that never says where it works gets flagged.
Title tag present and under 60 characters, exactly one H1, at least one internal link. Basic SEO hygiene the engines expect.
If you audited an ad in the same session, compares headline, offer, and phone number between the ad and the page. Mismatches kill conversions.
Paste your ad and up to 5 competitor ads or page texts you found yourself — automatic competitor discovery isn't included, and the page says so honestly. The same 62-pattern engine scores every entry, then derives the comparison. Run a benchmark →
Score, verdict, per-family violation counts, license presence, and trust-signal counts for you and each competitor — plus the competitor average row.
Your score against the competitor mean, and per-family wins and losses derived from the violation counts. Fewer violations than average is a win.
Your own findings grouped by violation family, HIGH first — with the same fix guidance as the ad audit. Derived from your copy only; nothing is invented about competitors.
Agency name field stamps the printed header for white-label client reports.
An AI audit can say something different every run and hallucinate evidence. A rule engine can't. Same ad in, same verdict out — and when it misses, the fix is legible: add a pattern, add a regression test, bump the version. Every report is stamped with the policy version it was judged against (currently v—).
MOCK PREVIEW — styled mock of a report layout, not a real audit.
"Nobody beats our prices — that's a promise."
How to fix: State the actual terms instead of absolute language: timeframe, conditions, and how the customer claims it.
"If you're like most homeowners, you're overpaying for energy."
How to fix: Describe the problem in general terms instead of the reader.
Pro Agency reports print clean — no tool chrome, just the verdict, the evidence, and the fixes — ready to hand to a client with your branding. Every report carries the policy version stamp, so a client can see exactly which rule book their ad was judged against.
See agency pricing