AI Transparency Notice

SemlyPro

Operator / imprint. This notice is published by Semly Pro, a sole proprietorship (eenmanszaak) established in the Netherlands, registered with the KvK (Dutch Chamber of Commerce) under number 99448351, VAT ID NL005387029B31, registered address Hawaiiweg 41, 1339 NW Almere, Netherlands (trading as “SemlyPro”, “we”, “us”, “our”). SemlyPro currently operates through this sole proprietorship; its sole proprietor is Surya Pillai. SemlyPro intends to incorporate a private limited company (Semly Pro B.V.). Upon incorporation, Semly Pro B.V. will assume this notice, and references to “SemlyPro” will be read as references to Semly Pro B.V.

Effective date: 13 July 2026 Location: semlypro.com/ai-transparency

1. What this notice is for

This AI Transparency Notice explains how SemlyPro uses AI, which AI models power the SemlyPro service, and how we implement the transparency and content-marking obligations under Article 50 of the EU AI Act.

It applies to all users of SemlyPro’s AI features, including:

• customers with a paid subscription or an Enterprise agreement; and
• users of the free AI-powered SEO tools on our marketing site (semlypro.com), who may have no account and no Data Processing Agreement with us.

Where this notice refers to how we notify “customers” of changes (for example, changes to our AI providers), free-tool users without an account are notified through the updated version of this notice and the relevant page on our site rather than by individual message.

This notice sits alongside our Terms of Service, Privacy Policy, Acceptable Use Policy, Cookie Policy, Data Processing Agreement, and Trust and Safety page. The legal entity that operates SemlyPro and publishes this notice is identified in the Operator / imprint block above and in Section 13.

1A. Territorial scope (EU, UK, India and worldwide)

SemlyPro is established in the Netherlands and serves users in the European Union, the United Kingdom, India and elsewhere. The transparency and marking practices in this notice are driven primarily by the EU AI Act, and we apply them by default to all users worldwide. In addition:

• European Union / EEA. The EU AI Act (Regulation (EU) 2024/1689) is the primary framework governing our AI transparency and content-marking obligations (Sections 2–10).
• United Kingdom. As of the last update of this notice, the UK has no direct statutory equivalent to the EU AI Act’s content-marking regime; the UK follows a principles-based, regulator-led approach to AI. However, UK consumer-protection and advertising rules — including the Digital Markets, Competition and Consumers Act 2024 (enforced by the CMA) and the ASA/CAP advertising codes — require AI-related disclosures not to be misleading. We therefore apply the same marking and disclosure practices described in this notice to UK users voluntarily. UK data-protection matters (UK GDPR, PECR) are addressed in our Privacy Policy and Cookie Policy, not here.
• India. India regulates AI-generated content through the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026 (notified 10 February 2026, in force from 20 February 2026), which impose labelling and permanent-metadata obligations for “synthetically generated information” (SGI). Our approach to those obligations is set out in Section 5.5. India data-protection matters under the Digital Personal Data Protection Act, 2023 (DPDP Act) are addressed in our Privacy Policy and our India Privacy & Grievance Addendum, not here.

2. Our position under the EU AI Act

For the purposes of Regulation (EU) 2024/1689 (the “EU AI Act”):

• SemlyPro deploys AI systems that are built on general-purpose AI models supplied by our third-party AI providers. Under the Act, actors deploy AI systems rather than deploying general-purpose AI models directly.
• In respect of the Content-generation and Content-analysis features we place on the market under our own brand, SemlyPro is a provider of an AI system within the meaning of Article 3(3) EU AI Act. Because those features generate synthetic text (and, where enabled, other synthetic content), our provider status triggers the machine-readable marking duty in Article 50(2) (see Section 5).
• Our customers are deployers of SemlyPro in their own capacity when they use it (Article 3(4)).

Where our customers rebrand or resell SemlyPro Content under their own brand, they may in turn become providers or further deployers under the value-chain rules in Article 25 EU AI Act, with associated obligations. Article 25 reclassifies actors as providers of high-risk AI systems in defined circumstances. SemlyPro’s SEO content-generation features are not high-risk (we prohibit the Annex III high-risk use cases listed in Section 9), so Article 25 is not the basis of our own provider status; we would only fall within Article 25 in respect of any future feature that is itself high-risk.

3. Which AI models we use

SemlyPro uses general-purpose AI models supplied by third-party AI providers to generate and analyse Content. The current list of AI providers is:

We update this list when we add, replace, or remove an AI provider. Changes are notified to subscribed customers in advance in accordance with our Data Processing Agreement, and reflected in the updated version of this notice for free-tool users. The authoritative, always-current list of our processors and sub-processors is published at semlypro.com/subprocessors, and this notice should be read consistently with it.

We do not train these models ourselves. SemlyPro is not a provider of a general-purpose AI model under Article 3(63) EU AI Act. We integrate models supplied by third parties.

Model versions change over time. AI providers release new versions, deprecate old ones, and update models continuously. Content generated by the Service may reflect model behaviour, biases, or limitations that vary across model versions and updates. This is documented in our Terms of Service, clause 8.4 (model changes).

3A. AI systems used for brand and AI-visibility tracking

Separately from Content generation, SemlyPro offers brand / AI-visibility tracking, which measures whether and how a brand is cited by AI search and answer engines (for example ChatGPT, Claude, Perplexity, Google Gemini and Google AI Overview / AI Mode). Where this feature sends queries to those third-party AI systems, it uses read-only measurement queries and does not submit your Customer Materials to them for training or generation. To the extent any of those systems act as recipients or sub-processors of personal data, they are disclosed in our Data Processing Agreement and the sub-processor list at semlypro.com/subprocessors, which govern that processing. Section 3 (Content generation) and the sub-processor list are intended to be consistent; where they differ, the sub-processor list is authoritative.

4. How Content is generated

At a high level, SemlyPro generates Content as follows:

• Grounding. For content-generation requests, we scrape and analyse the top-ranking pages for the target keyword, extract structural signals (headings, entities, word count, intent), and build a structured content brief. Where you have connected data sources, the brief may also draw on data you bring into the Service — including Google Ads (campaign, spend, performance and conversion data), Google Analytics 4 (traffic, audience and event data), Google Search Console (query, impression, click and index data), and files you upload (including Excel/CSV) containing arbitrary tabular data. This connected and uploaded data is treated as Customer Materials under our Terms of Service and, where it contains personal data, is processed under our Data Processing Agreement, with you acting as controller/data fiduciary. Our use of the Google integrations is further described in our Google API Services / Limited Use Compliance Statement and Privacy Policy.
• Generation. The content brief is sent to an AI provider’s model together with your Customer Materials (brand data, tone, target audience). The model generates a draft.
• Post-processing. We apply structural checks, format conversion for the destination CMS, and — as our content-marking implementation goes live (see Section 5) — machine-readable content marking.
• Delivery. The Content is returned to you for review and editing.

We do not fact-check Content. AI models may produce inaccuracies, fabricated statistics, invented product features, or biased framings. You must review Content before publishing. Full detail in Terms of Service, clause 8 (AI output — no warranty; your responsibility to review).

5. Article 50 content marking

5.1. When marking applies

Under Article 50(2) EU AI Act, providers of AI systems that generate synthetic content must mark that content in a machine-readable format detectable as artificially generated or manipulated, using solutions that are effective, interoperable, robust, and reliable insofar as technically feasible.

The dates that apply to SemlyPro are:

• From 2 August 2026 for any SemlyPro AI system first placed on the market on or after 2 August 2026 — marking is required from placement.
• From 2 December 2026 for any SemlyPro AI system already on the market before 2 August 2026 — under the EU’s Digital Omnibus amendments to the AI Act, generative AI systems already on the market before 2 August 2026 have until 2 December 2026 to meet the Article 50(2) machine-readable-marking requirement. The general Article 50 disclosure duties (Sections 6 and following) are unaffected by this transitional window.

Marking earlier than the applicable date is permitted and protective, and we intend to mark by default (see Section 5.4). The exact wording of the transitional provision should be confirmed by counsel against the final text of the Digital Omnibus as published in the Official Journal of the EU.

5.2. How we mark

We are building our content-marking implementation to satisfy Article 50(2). The mechanisms below describe our intended implementation; each must be verified against the actual production build before this notice is published, and any mechanism not yet shipped will be described as a forward-looking commitment (with a target date) or removed.

For text output, we are implementing content marking via:

• metadata markers embedded in the output alongside content-provenance information (designed to be compatible with the C2PA — Coalition for Content Provenance and Authenticity — content credentials specification where technically feasible);
• HTML/JSON structured markers in publishable outputs, including <meta> tags identifying the AI provider and generation timestamp; and
• response-header markers in API responses, using X-AI-Generated: true and related headers.

For image output (a feature that is conditional — only where SemlyPro adds image generation): C2PA content credentials embedded in image metadata.

For MCP outputs: markers in the MCP response envelope.

Detailed technical documentation is intended to be published at semlypro.com/ai-transparency/technical.

5.3. Limitations of marking

Content marking is a defensive layer, not a guarantee. Determined actors can strip markers from AI-generated text more easily than from other media types. We do not provide any feature designed to remove or suppress our markers, and — as described in Section 5.5 — for content within scope of India’s SGI rules we aim to embed persistent provenance metadata; but we cannot prevent third parties from attempting to alter or remove markers after Content leaves the Service. We monitor and improve our implementation as detection technology and Article 50 Code-of-Practice guidance evolve.

5.4. Provider marking exceptions

Article 50(2) EU AI Act includes narrow provider marking exceptions: where the AI system performs an assistive function for standard editing, or does not substantially alter the input data or its semantics; and for AI systems authorised by law to detect, prevent, investigate or prosecute criminal offences. These exceptions do not typically apply to commercial SEO content generation, so in practice we mark by default once our implementation is live.

(The separate artistic, satirical, fictional and editorial-responsibility carve-outs are deployer exceptions under Article 50(4), not provider marking exceptions; they are addressed in Section 6.)

5.5. India — synthetically generated information (SGI) labelling

For users and content within scope of India’s Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026 (in force from 20 February 2026), AI-generated Content produced through SemlyPro is “synthetically generated information” (SGI). Those Rules expect a computer resource that enables the creation or modification of SGI to ensure such content is prominently labelled as synthetic/AI-generated and embedded with a permanent metadata or other technical mechanism identifying that it was created using the resource, and that the label and metadata are not made removable through the tool itself.

Our approach is to: (a) label AI-generated Content as synthetic/AI-generated; (b) embed provenance metadata designed to be as persistent as technically feasible; and © not provide any feature that strips or suppresses that label or metadata. As noted in Section 5.3, text markers are inherently more fragile than image markers, and we cannot guarantee that third parties will not attempt to remove them downstream after publication. Data-protection obligations for Indian users under the DPDP Act, 2023 are addressed in our Privacy Policy and India Privacy & Grievance Addendum.

6. Your deployer obligations (from 2 August 2026)

When you publish Content generated through SemlyPro, you are a deployer under Article 50(4) EU AI Act. Depending on what you publish, you must:

• For text published to inform the public on matters of public interest: disclose that the content is artificially generated or manipulated, save for the exception in Article 50(4) where the content has undergone a process of human review or editorial control and where a natural or legal person holds editorial responsibility for the publication. This is the live obligation for SemlyPro’s current text output.
• For “deep fakes” (AI-generated or manipulated image, audio or video content, as defined in Article 3(60) EU AI Act): disclose clearly and distinguishably that the content is artificially generated or manipulated, save for the narrow artistic/satirical/fictional and analogous exceptions in Article 50(4). This obligation applies only if and when you use image, audio or video generation. SemlyPro currently generates text; image generation is conditional (see Section 5.2), and audio/video generation is not offered.

We provide tools in the SemlyPro user interface that make these disclosures straightforward, including:

• a disclosure checkbox at publish time;
• standard-language disclosure snippets your editor can insert; and
• machine-readable markers automatically embedded in the output (as our marking implementation goes live — see Section 5).

7. Serious-incident reporting (voluntary; Article 73 reference)

Article 73 EU AI Act imposes mandatory serious-incident reporting on providers of high-risk AI systems (and, via Article 55, on general-purpose AI models with systemic risk). SemlyPro is neither: it prohibits the Annex III high-risk use cases (Section 9) and is not a provider of a general-purpose AI model (Section 3). Article 73 therefore does not strictly apply to SemlyPro’s current SEO content-generation features.

Nonetheless, as a voluntary commitment, from the effective date of this notice SemlyPro will report serious incidents involving the Service to the competent market surveillance authority, taking the Article 73 timeframes as our reference where relevant:

• Standard serious incidents: we will report without undue delay and, in any event, no later than 15 days after we become aware of the incident and have established a causal (or reasonably likely causal) link to the Service.
• Widespread infringement or a serious and irreversible disruption of critical infrastructure (Article 3(49)(b)): no later than 2 days after we become aware.
• Death of a person: no later than 10 days after we become aware.

If we ever offer a feature that is itself high-risk, Article 73 would apply to that feature as a matter of law and we would report accordingly.

Customers must notify SemlyPro of any serious incident within 24 hours of becoming aware, by writing to anil@semlypro.com. Details in Terms of Service, clause 8A.5 (serious-incident notification).

8. Voluntary Code of Practice

We monitor and, where appropriate, align with the EU AI Office’s Code of Practice on Article 50 as it develops. The current position of that Code (as of the last update of this notice) is published at digital-strategy.ec.europa.eu.

9. Prohibited practices and high-risk uses

SemlyPro prohibits use of the Service for:

• any practice prohibited by Article 5 EU AI Act; and
• any Annex III high-risk use case (employment/hiring, credit scoring, education, essential services, law enforcement, migration, justice, biometric identification) without our prior written consent.

The full list and the mechanism for requesting written consent are in our Acceptable Use Policy (see AUP section 3.6, regulated sectors and YMYL, and the EU AI Act prohibited-practices and high-risk provisions in that policy).

10. Human oversight

SemlyPro is designed for human-in-the-loop use. Every Content item generated by the Service is intended to be reviewed by a human before publication. Our editor UI, our publishing workflow, and our documentation all reflect this design.

We do not currently support fully autonomous publication for high-risk topics (see the regulated-sectors provisions in our Acceptable Use Policy, section 3.6).

11. How we use your data (no training of public models; bounded staff access)

No training of public models. SemlyPro does not use Customer Materials or Content generated for you to train publicly-released AI models. Our contracts with each AI provider require the same on their side by default.

Bounded internal access and improvement. So that we can run the Service, SemlyPro personnel and authorised contractors may access Customer Materials and Content only as necessary to (a) provide, maintain and secure the Service; (b) troubleshoot, debug and provide support; © investigate abuse, security or Acceptable-Use issues; and (d) review and improve the Service and its models for SemlyPro’s internal, non-public purposes only. Such access is subject to confidentiality obligations, least-privilege/role-based access controls, and access logging. To be clear, this internal, non-public review and improvement is not the same as training publicly-released AI models, which we do not do.

Full details are in our Privacy Policy section 5 (how we use data / AI processing) and our Data Processing Agreement.

12. Changes to this notice

We update this notice when:

• we add, replace, or remove an AI provider;
• our content-marking implementation materially changes;
• Article 50 Code of Practice guidance materially changes;
• India SGI-labelling requirements or other applicable AI-transparency laws materially change; or
• we introduce a new AI feature with new transparency implications.

The “Last updated” date at the top of this notice reflects the most recent revision.

13. Contact

• Questions about AI transparency: anil@semlypro.com
• Report an AI-related concern (including model behaviour, hallucination or quality issues): anil@semlypro.com
• Report a serious AI incident: anil@semlypro.com
• India — Grievance Officer (for Indian users’ grievances, including AI-content concerns): anil@semlypro.com (see our India Privacy & Grievance Addendum for the Grievance Officer’s full details and response timelines).

Operator. This notice is published by Semly Pro (eenmanszaak), trading as SemlyPro, KvK 99448351, VAT NL005387029B31, Hawaiiweg 41, 1339 NW Almere, Netherlands — to be succeeded by Semly Pro B.V. on incorporation, which will assume this notice.

Last updated 13 July 2026 · Operated by Semly Pro (eenmanszaak), KvK 99448351, Hawaiiweg 41, 1339 NW Almere, Netherlands · Questions about this document: anil@semlypro.com