AI Call Disclosure | Airis

Last Updated: May 13, 2026

1. Clear AI disclosure

Airis uses AI to interact with callers and collect job details for trades businesses. Callers must be told that they are interacting with an AI assistant, unless that is already obvious in the context.

2. Standard opening line

Our required pilot opening line is:

Airis should not pretend to be a human receptionist. Businesses should not configure scripts or public messaging that hide the use of AI.

3. Scope of the assistant

The assistant is designed to capture the caller's name, phone number, job/problem description, address or Eircode, urgency and callback details, then pass those details to the business. It is not designed to diagnose work, quote prices, promise attendance, or provide professional safety advice.

4. AI limitations

AI may mishear, omit, summarise poorly or classify urgency incorrectly. The business/operator remains responsible for checking the details, contacting the caller, deciding whether and how to respond, and making any professional or safety decisions.

5. Emergency and safety wording

Airis is not an emergency service and does not guarantee response times. If there is immediate danger to life or health, fire, gas, electricity or another serious emergency, the caller should contact emergency services or the relevant utility/emergency provider.

6. Recording, transcription and summaries

Airis processes live call audio through telephony and AI providers to provide the service. The application stores transcript text and AI summaries where technically produced, but the repo does not show intentional raw call audio storage in Airis's own database. Provider-side audio, transcript or log retention must be confirmed in the relevant Twilio/OpenAI/account settings before real caller pilots. If a business enables or requests call recording, or uses Airis in a way that requires additional notices, the business is responsible for providing those notices to callers.