AI Transparency Statement

Last updated: 6 July 2026 · Prepared for the EU AI Act transparency obligations (Regulation (EU) 2024/1689, Article 50, applying from 2 August 2026)

We build AI systems for a living, so we hold ourselves to the standard we'd design for a client. This page says plainly where AI is used on this site, what it may and may not do, and where the humans sit.

You are talking to an AI

The assistant in our client workspace — Ray — is an artificial intelligence system operated by Quantum Touch Limited (trading as Strange Materials). It is labelled as AI throughout the product, and this page is the formal declaration: when you chat with Ray, by text or voice, you are interacting with a machine, not a person. If Ray's voice mode is on, the voice you hear is synthetic speech generated by AI.

You can ask for a human at any time — say so in the chat and the request reaches our team directly. Quotes, prices and commercial commitments are never sent by the AI: they are prepared, reviewed and issued by people.

What Ray does

• Holds scoping conversations about your business problem and asks the questions a consultant would ask.
• Reads documents you choose to upload (PRDs, spreadsheets, presentations) and, where your email address belongs to a business domain, reads your public website so it arrives prepared.
• Drafts working documents: an opportunity brief, a business case (as an editable Excel model), a pitch deck and a draft scope. Every one is marked as an AI-assisted draft and states that it is not a proposal — our formal proposal follows separately and is the only document that binds.
• Labels its numbers. Figures in our documents carry their source — stated by you, a labelled public fact, our labelled estimate, or calculated — and the business case is built so you can change the assumptions yourself.

What Ray is built on

Ray runs on Claude large language models provided by Anthropic, accessed via API. Anthropic does not use our API data to train its models. Anthropic is the provider of the general-purpose AI model; Quantum Touch Limited is the provider and deployer of the Ray system built on top of it. Data handling, including international transfer safeguards, is covered in our Privacy Policy; where material you provide contains other people’s personal data, our Default DPA governs that processing.

Risk classification and oversight

• Classification. Ray is not a high-risk AI system under Annex III of the EU AI Act — it does not make decisions about employment, credit, essential services, law enforcement or any similarly protected area. It is an interactive assistant subject to the Act's transparency obligations, which this page and the in-product notices fulfil.
• Human oversight. Every commercial output passes a human gate: our team reviews and issues all quotes and proposals; an automated review layer additionally screens AI-drafted document packs and holds anything questionable for human review before it reaches you. A decision to decline a project is made by a person, never by the AI.
• No automated legal effects. No decision producing legal or similarly significant effects about you is made solely by automated means (GDPR Article 22). The AI drafts; people decide.
• Content marking. AI-assisted documents carry human-readable notices on their pages and machine-readable marking in their file metadata.
• Staff AI literacy. The people who build and operate our systems work with AI daily and are trained in its capabilities, limits and failure modes, in line with Article 4 of the EU AI Act.

Limits — read this part

Ray's outputs are estimates and drafts, produced to speed up a conversation, not to end one. AI systems can be confidently wrong. Figures are labelled with their sources so you can check them, and the formal review happens on our side before anything binding is issued — but do not make financial, legal or investment decisions on an AI draft alone. Our documents are not legal, financial, tax or professional advice.

Where this applies

We apply the strictest standard we operate under — the EU AI Act together with GDPR — to all visitors, wherever you are. If you're reading from outside the EU, you get the same disclosures and the same human oversight.

Questions, complaints, incidents

Anything about our use of AI — including a request to have a human review something the AI produced — goes to [email protected]. If something an AI system of ours did seems harmful or seriously wrong, tell us and we will investigate promptly. You can also raise data-protection concerns with the Irish Data Protection Commission (dataprotection.ie).