This policy explains how AutonomicsAI LLP ("Autonomics AI", "we") uses artificial intelligence (AI) in its work, and sets the rules for what we will and will not do for clients. It forms part of our Terms of Service for Business Clients. If this policy and a Statement of Work conflict, the Terms of Service decide which prevails.
Part A: How we use AI
A1. Where we use it. We use AI tools to help research, write, design, code, analyse data, draft emails and replies, prepare audits and process catalog data. We also build automations for clients that may use AI, such as instant replies to enquiries.
A2. Human review. A person on our team reviews AI output before we deliver it to a client or send it to anyone on our own behalf. For automations that run without a person in the loop (for example an instant reply sent within seconds), the client approves the message templates, rules and limits before launch, and we test them before they go live.
A3. You review before publishing. AI can be wrong, out of date or biased, and can produce text that sounds confident but is false. You must check all content we prepare before you publish, send or rely on it, especially prices, licences, insurance, warranties, safety information, specifications and legal or regulated claims.
A4. Your data and AI training. We do not use your confidential information or your customers' personal data to train AI models. We use AI tools on business accounts that do not allow our data to be used to train models. We do not put client personal data into consumer AI accounts. We will only train or fine-tune a model on your data if you ask for it in writing and agree the terms.
A5. Tools that touch client data. The AI tools we use to process client personal data are described in our Data Processing Agreement as a category of sub-processor, with their processing locations, and we tell clients before adding a new kind of tool.
A6. Disclosure. We tell clients that our delivery is AI-assisted (this policy does that). Where the law requires people to be told that they are interacting with an AI system, or that content is AI-generated or manipulated, the systems we build will make that disclosure, and we will not remove it. This includes the transparency rules in Article 50 of the EU Artificial Intelligence Act, which apply from 2 August 2026, where they apply.
A7. Ownership. Content made with AI tools may not be protected by copyright, and similar content may be produced for others. Our Terms of Service explain ownership.
A8. Accuracy of AI-driven numbers. Audit scores, estimates and forecasts produced with AI are estimates. We say so in the audit, and we do not present them as guarantees.
Part B: What we will not do
We will not create, send, publish or operate any of the following, for a client or for ourselves, and clients must not use our services or Deliverables to do so:
B1. Fake or manipulated reviews and endorsements. Writing or generating reviews or testimonials that do not reflect a real customer's real experience; buying or selling reviews; reviews by owners or staff presented as customer reviews; offering incentives conditional on positive reviews; suppressing negative reviews; or fake social-media followers or engagement. (In the US, these practices are prohibited by the Federal Trade Commission's Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465.) We will set up review requests that go to all customers, without filtering by sentiment.
B2. Spam and unlawful messaging. Sending marketing emails, SMS or messages without the consent or other lawful basis that the recipient's law requires; using bought or scraped personal lists; hiding who is sending a message; false or misleading headers or subject lines; ignoring opt-outs; or messaging people on a suppression list. We will not send cold email to countries whose law requires prior consent for business email (for example Germany and Austria), or to UK sole traders and partnerships without consent.
B3. Deceptive claims. False or unsubstantiated claims about prices, discounts, licences, insurance, certifications, guarantees, "compliance", results, health or safety; fake urgency or scarcity; fake locations, fake "local" presence or fake business listings; or impersonating another business or a real person.
B4. Deceptive AI use. AI-generated people, voices or images presented as real customers, staff or events; deepfakes of real people; or chatbots that claim to be human when sincerely asked.
B5. Infringement. Copying competitors' content, images or trade marks; using images or fonts without a licence; or keyword or ad tactics that misuse another business's trade marks unlawfully.
B6. Unlawful data practices. Collecting personal data without the notice or lawful basis the law requires; collecting more than is needed; covert tracking; or processing sensitive data (such as health data) without the safeguards the law requires.
B7. Accessibility overlays and claims. Claiming that a website is "ADA compliant", "WCAG compliant" or "fully accessible" because of our work, or relying on overlay widgets as a substitute for fixing the underlying code.
B8. Platform abuse. Breaking the rules of Google, Meta, LinkedIn, marketplaces or other platforms, including fake accounts, automated activity that their terms forbid, and attempts to evade their reviews or bans.
B9. Illegal or harmful businesses. Work for businesses or offers that are illegal where they operate or sell, or for scams, weapons trafficking, adult content, gambling without a licence, or products that make unlawful health claims.
Part C: If a request crosses the line
C1. If we believe a request is covered by Part B, we will tell you why and suggest a lawful alternative where one exists.
C2. If you insist, or if we find that a Deliverable is being used in breach of this policy, we may refuse the work, take down content or automations that we control, and suspend or end the services under our Terms of Service. Fees for work already done remain payable.
C3. You are responsible for content you approve and for the way you use the Deliverables, and you will indemnify us as set out in our Terms of Service.
Part D: Reporting a concern
To report misuse of something we built, or to raise a concern about our use of AI, email [email protected]. We review every report.