This policy discloses how Toutmark uses AI and commits to transparent, honest use of generative models. Toutmark complies with FTC guidance on AI disclosure (FTC Section 5 of the FTC Act, the AI Comply rule, and the Endorsement Guides) and does not use AI to deceive.
0. AI Disclosure Mandate (binding for every Toutmark customer)
Every customer-facing surface where Toutmark generates or substantially edits content carries a visible, machine-readable AI disclosure. This applies to ALL Toutmark customers regardless of plan tier, regulated/non-regulated status, or approval mode. There are no carve-outs. The mandate exists because uniformity is the legal defense — once any customer's content is undisclosed, regulators can argue the rest is deceptive too.
Where the disclosure appears
- On-site customer pages (paragraph rewrites, Q&A pages, education pages, comparison and rankings pages): visible footer or page-level note — "Some content on this page was drafted with AI assistance and reviewed for accuracy by [Customer Name]."
- Q&A / FAQ pages: same footer plus
"creativeWorkStatus": "AI-assisted, human-reviewed"field in the FAQPage JSON-LD schema (machine-readable). - HARO pitches: no AI disclosure in the pitch body (most source platforms — HARO, Connectively, Featured, Qwoted — explicitly reject AI-disclosed pitches in their TOS). Human accountability is enforced through Pangram humanizer pre-flight, customer review + approval before send, and Spokesperson Attestation in our Terms of Service. The pitch ends with attribution only: "— [Spokesperson Name], [Spokesperson Title], [Customer]".
- EIN press releases: footer line below the boilerplate — "This release was prepared with AI assistance and reviewed by [Approver Name + Title] before distribution."
- Wikidata edits: covered by the separate Wikidata bot-account disclosure page; bot edits are tagged as such in the platform's edit history.
- Wikipedia edits: covered by mandatory WP:PAID disclosure on the editor's user page.
- Toutmark's own marketing (blog posts, audit reports, cold emails): disclosure on every piece. Cold emails carry a short P.S. line — "This email was personalized using AI."
What we don't do
We do not hide the disclosure via CSS, opacity, aria-hidden, or off-screen positioning. We do not use a font size below the page's baseline body text size. If a customer's CMS template strips or hides the disclosure, our Monitor agent flags it and our Webmaster re-injects it; three consecutive removals is grounds for offboarding the customer with refund.
Customer responsibility
By signing up for Toutmark, the customer agrees the disclosure will appear on every applicable surface and represents to Toutmark that they have authority and approval (from the customer's organization, not merely an individual employee) for that disclosure to appear. Customers requesting disclosure removal at any time during the engagement will be offboarded; this clause is non-negotiable.
0a. Comparative Content Policy (Layered)
Toutmark's comparative-content policy stacks two layers. Layer 1 applies to all customers regardless of industry. Layer 2 differs by industry tier.
Layer 1 — No Direct-Competitor Comparisons (all customers)
Toutmark will not name, identify, rank, criticize, or make comparative claims about your direct competitors in any AI-assisted content drafted, edited, submitted, or published on your behalf. This applies to HARO pitches, EIN press releases, on-site paragraph rewrites, Q&A pages, education pages, rankings pages, blog posts, social posts, llms.txt prose, schema text fields, and every other customer-facing surface. AI-generated comparative claims about named competitors are the highest-risk content category for trade-libel, Lanham Act § 43(a) false-advertising, state UDAP/Little FTC Act, FTC Section 5, and tortious-interference exposure.
Rare opt-in path. Where a customer specifically needs named-competitor content for a public statement, we may proceed only with all three of: (a) the customer's own attorney's pre-clearance of the specific claim with documented substantiation; (b) the customer's personal verification of the claim's accuracy; and (c) execution of a per-claim Comparative Claim Authorization in the customer's dashboard, capturing typed verification phrase, attorney-name-on-file, timestamp, and IP. Even with all three, Toutmark's Editor reserves the right to decline if substantiation is not ironclad. The Editor's decline is final.
Layer 2 — Comparison-of-Any-Kind Rule (tier-aware)
If you are in a regulated industry (registered investment advisor, broker-dealer, hedge fund or PE/private credit, healthcare provider or HIPAA covered entity, attorney, insurance producer, lender or mortgage, crypto/digital-asset, real estate broker, telehealth, MLM, tax/CPA), Toutmark will not include ANY comparative claims in AI-assisted content drafted on your behalf — including category-level comparisons ("better than typical CRMs"), industry-baseline comparisons ("above the industry average"), customer-own-past-performance comparisons ("30% improvement vs. last quarter"), or legacy-versus-modern framing ("unlike traditional approaches"). The blanket no-comparisons rule maps to your regulator's specific restrictions on comparative communications: SEC Marketing Rule 206(4)-1, FINRA Rule 2210, state medical board and FTC healthcare advertising rules, state bar attorney advertising rules (including ABA Model Rule 7.1), state insurance commissioner rules, and analogous regimes. Toutmark's flat-zero rule for regulated customers exists to preserve your regulator-compliance posture; it cannot be opted out of.
If you are not in a regulated industry, comparisons are permitted in AI-assisted content Toutmark drafts on your behalf — at the category level, industry-baseline level, customer-own-past-performance level, or industry-trend-observation level. Every factual claim within a comparison must be source-verifiable: a public study, a published industry report, a customer-supplied internal data point with documented methodology, a published benchmark, or a similarly citable source. Toutmark's Editor will not include hallucinated facts; comparative claims that cannot be verified against a source will be rejected. You may supply your own benchmark data at onboarding or per-claim, or rely on Toutmark to source public benchmarks with citation.
Cherry-picking guard. Even with verified facts, Toutmark's Editor reserves the right to reject a comparative draft if the comparison is correct on a single dimension but materially misleading in context (e.g., the customer's product is faster but materially worse on accuracy or reliability that buyers would care about).
1. AI Systems We Use
Content LLM (drafting + Toutmark agent reasoning)
Toutmark uses one LLM for all content drafting and internal agent reasoning:
- Anthropic Claude (Sonnet 4.6 default; Haiku 4.5 fallback if Sonnet is unavailable): Single content-and-reasoning LLM. Used for every paragraph rewrite, Q&A page, education page, HARO pitch, EIN press release, blog post, audit report, schema text field, llms.txt prose blurb, and every Toutmark agent's internal reasoning. Toutmark does NOT use OpenAI, Google Gemini, or Perplexity for content drafting — they are measurement targets only (see below).
AI Citation Measurement Targets (third-party answer engines we query to score customer brand visibility)
Core to Toutmark's AI citation product is measuring whether customers' brands surface in third-party AI assistants when users ask category-relevant questions. To do this, our Monitor agent periodically queries each of the major answer engines with brand-test prompts and parses the responses to determine citation frequency. The following services are queried for measurement only — never for content drafting:
- OpenAI (GPT-5-mini and text-embedding-3-small only): Citation-measurement queries against ChatGPT's underlying models, plus fact-verification embeddings supporting the comparative-claim Editor gate.
- Google Gemini (Gemini Flash): Citation-measurement queries against Google's answer engine.
- Perplexity: Citation-measurement queries against Perplexity's answer engine.
Email Verification
Toutmark uses NeverBounce to verify recipient email-address deliverability before sending cold outreach and to filter invalid, role-based, or disposable addresses on prospect intake. Only the email address itself is sent to NeverBounce — no message content, no other personally identifying information.
AI-Text Detection
Toutmark uses Pangram Labs to evaluate AI-generated drafts before they are published or sent. If a draft scores too clearly as machine-written, our humanizer rewrites and re-scores it; drafts that still fail are rejected rather than shipped. Only the generated text is sent to Pangram — no PII and no account data.
AI Applications
Toutmark applies AI to:
- Draft and rewrite website content optimized for search engines and AI assistants
- Generate social media posts, email sequences, and brand-aligned copy
- Analyze competitor content and identify optimization opportunities
- Score and monitor how AI assistants cite and recommend your business
2. Transparency & Disclosure
AI Disclosure in Marketing
Toutmark discloses AI use transparently:
- Website: Our homepage and service pages clearly state that Toutmark is AI-powered
- Onboarding: New customers receive onboarding material explaining how Toutmark uses AI
- Cold Outreach: All cold emails from Toutmark disclose AI use within the first 3 sentences
- Social Media: Posts clearly state when AI was involved in content creation
No Deceptive AI Use
Toutmark does NOT:
- Pretend AI-generated content is human-written
- Hide AI involvement or use dark patterns
- Claim AI can guarantee specific results
- Use AI to create deepfakes or voice clones
- Generate fake testimonials or fabricated reviews
2a. Regulated Industries — Compliance Scope Disclosure
Toutmark provides content optimization tooling, not legal or regulatory compliance advice. Customers operating in regulated industries are subject to additional advertising, disclosure, and record-keeping rules that Toutmark's default mode does not enforce. We classify regulated customers into three tiers based on the dominant regulator and the strength of the testimonial-rule constraint that applies.
Tier 1 — full Regulated Industry Mode required, no exceptions: securities and investment advisory (RIAs, ERAs, broker-dealers, private funds — SEC Marketing Rule 206(4)-1, FINRA, state blue-sky laws); healthcare and HIPAA covered entities (hospitals, clinics, dental, mental health, pharma, medical devices, supplements making health claims — HIPAA + FDA + FTC); legal services (law firms, solo attorneys — state bar advertising rules); insurance (agents, brokers, carriers — state insurance commissioner rules + NAIC Model Rule); cannabis (federal illegality + state-by-state advertising); gambling, sports betting, daily fantasy (state licensing + age-gating + problem-gambling disclosures); adult content; firearms and weapons; tobacco, vaping, and nicotine products. Toutmark does not offer review-site solicitation or response work, blog auto-publish, or cold email campaigns to Tier 1 customers under any tier or pricing arrangement.
Tier 2 — restricted service menu with compliance review on every draft: financial services that are not technically securities (banks, lenders, fintech, payments, BNPL, debt relief, credit repair — CFPB UDAAP, Reg Z/TILA, CROA, Telemarketing Sales Rule); crypto, web3, and digital assets (SEC/CFTC uncertainty, state money-transmitter rules, post-Kardashian celebrity-endorsement enforcement); real estate brokerages and agents (RESPA Section 8, Fair Housing Act, state Realtor codes of ethics); telehealth and DTC pharma (Ryan Haight Act, FDA promotional rules); MLM and direct sales (FTC Business Opportunity Rule, state earnings-claim rules); tax preparation, accounting, and CPAs (IRS Circular 230, state CPA board advertising rules). Tier 2 customers receive Regulated Industry Mode by default; review-site work is generally still excluded; AI citation work runs with Editor compliance review on every rewrite.
Tier 3 — full standard menu with one specific service-surface restriction: EdTech, childcare, and any business serving under-13 audiences (COPPA + FERPA on data collection — Toutmark does not touch your data systems, marketing-side AI citation work is fully fine); political campaigns and advocacy (FEC + state campaign-finance disclosure — Toutmark declines these on neutrality grounds); nonprofits (state charitable-solicitation registration on fundraising-adjacent advertising — Toutmark accepts but does not run donor-targeted cold email).
For regulated-industry customers we offer Regulated Industry Mode, which adds the following constraints to all Toutmark services:
- Stricter content fidelity: disclaimers, qualifiers, and risk disclosures in your original content remain in the same paragraph as the claim they modify; they are not repositioned, de-emphasized, or moved to footers during paragraph restructuring.
- No auto-publish: every draft rewrite, HARO response, press release, Wikidata edit, and Wikipedia draft routes to your compliance team's review queue in the dashboard. Your compliance team approves before anything ships publicly. Toutmark's default Editor-autonomy doctrine does not apply.
- Excluded services: G2 / Capterra review solicitation, blog post auto-drafting with auto-publish, and any cold email campaigns run on your behalf are not offered to regulated-industry customers, regardless of plan tier.
- HARO + press topical restrictions: HARO pitches are restricted to market-commentary / industry-trends / leadership-thought content; we do not draft performance claims, deal-specific, fund-specific, or solicitation content. Press releases are restricted to non-offering news (hires, awards, non-offering milestones).
- Wikidata scope limit: basic identity data only (name, founding year, location, leadership). Performance metrics, AUM, fund-level data, or specific deal outcomes are not added to Wikidata on your behalf.
- Audit log: every edit, rewrite, draft, review, and publish action is logged with 7-year retention, covering SEC Rule 204-2 Books & Records requirements, FINRA record-keeping rules, and equivalent state-bar or healthcare record-keeping rules. You can export the log for regulatory examination or internal audit at any time.
Liability posture: Toutmark provides tools that make compliance workflows easier — we do not provide legal or compliance advice. Regulated Industry Mode reduces operational risk but does not eliminate it. Your internal compliance team and retained legal counsel remain the final arbiter of whether specific content meets your regulatory obligations. By using Toutmark in Regulated Industry Mode, you acknowledge that (a) you are responsible for your own regulatory compliance, (b) Toutmark's drafts are proposals subject to your review and approval, and (c) Toutmark is not liable for regulatory actions taken against content you approved and published.
2b. Securities Industry — Specific Service Scope
If your business is a Registered Investment Adviser, Exempt Reporting Adviser, broker-dealer, hedge fund, private equity firm, private credit firm, venture fund, or family office, the following clarifies what Toutmark's services do and do not include for you, in light of SEC Marketing Rule 206(4)-1, FINRA Rule 2210 (where applicable), Reg D 506(b) and 506(c), the Investment Advisers Act of 1940, the Investment Company Act of 1940, and applicable state blue-sky laws.
Services we offer securities customers:
- Paragraph restructuring of existing site content under our regulated-fidelity rule (disclaimers, qualifiers, and risk disclosures stay in place verbatim during paragraph restructuring).
- Schema markup (FAQ, About, Service, FinancialProduct, FinancialService) — no performance metrics, no testimonials in schema unless the customer's compliance team has signed off.
- Brand-presence pages on the customer's own domain (firm-profile, education, FAQ pages) — never an "AI profile" or "AI citation profile" page name.
- HARO and EIN Presswire pitch drafting — every pitch goes through customer-side compliance review before submission. Toutmark never submits without explicit per-pitch approval from the customer's compliance contact for regulated industries.
Services we do NOT offer securities customers:
- No performance claims, return projections, or backtested track records in any AI-generated copy.
- No solicitation language for unregistered offerings (Reg D 506(b)) — anything we draft for a private fund is general brand-presence content only.
- No testimonials or endorsements unless the customer's compliance team has approved the testimonial under the SEC Marketing Rule and provided the required disclosures.
- No comparisons to specific named competitors. For regulated customers we apply a hard "zero-comparison" rule.
- No content directed at retail investors for funds that are accredited-only or qualified-purchaser-only.
The Securities Visual Approval Gate in your dashboard requires explicit per-action sign-off from your designated compliance contact before any draft is queued for publish. This gate cannot be disabled.