AI transparency and data, in plain English.

One short line, then a natural conversation

Every conversation opens with a warm greeting in your business’s name that says it’s an AI assistant — for example “Hi, I’m Niamh, Riverside Dental’s AI assistant.” After that, it’s simply a natural conversation in your tone: no robotic menus and no repeated disclaimers.

That one line is there because the EU AI Act asks for it: AI systems that talk with people must make clear they’re an AI, unless that’s already obvious. The rule has applied since 2 August 2026 (Article 50). If a customer asks whether they’re talking to a person, it answers honestly. Work that happens behind the scenes — summaries and records for your team — doesn’t talk to anyone, so it needs no mention.

Anything that needs you comes straight to you

Each setup has handover rules agreed with you — emergencies, complaints, payments, anything sensitive, anything outside approved answers — and what the customer is told. Those conversations are passed to your team with a short summary, and if a customer asks to speak to someone, it takes their details so a person can get back to them. The step-by-step walkthrough shows where those rules apply and what you control.

It answers from what you approve

Each agent works from a knowledge base built from your public information and anything you send us. You review the critical parts — prices, hours, services and policies — before anything goes live, and you can change them at any time.

Text from websites, documents or customers is treated as information, never as instructions to the agent. That protects your setup from someone trying to talk it into doing something it shouldn’t.

Extra care in sensitive sectors

For clinics, physiotherapists, vets and other regulated services we keep agents deliberately narrow: general questions, appointment requests and handover. No diagnosis or clinical advice, and no health details beyond what a request needs.

Only the data the job needs

We agree with you what each agent collects — typically a name, contact details and what the customer needs — and nothing more. Agents don’t take card details or payments in conversation.

When we handle your customers’ data, we do it on your behalf and under your instructions. Before any real customer data is processed, we put a data processing agreement in place with you. It lists the service providers involved (for example the AI model and hosting providers), where they process data, and how long information is kept.

How long information is kept

Retention is agreed with you for each setup and written into the data processing agreement. When it ends, data is deleted or returned. You can ask us to delete a customer’s information at any time.

Security basics

• Access to conversations and leads is limited to the people who need it, with individual accounts.
• Passwords and API keys stay on servers, never in the browser.
• Data is encrypted in transit (HTTPS/TLS).
• Changes to leads and settings are logged.
• New connections are tested with sample data before they touch real customers.

What we don't claim

We don’t claim certifications we don’t hold. If your business needs a particular standard or contract term, ask us — we’ll tell you plainly where we stand.