Your accounts. Your data.
Connecting AI to your business is a real decision. Here, in plain English, is what we can see, what we never see, and the rules we follow.
The rules
We never ask for passwords
Every connection is approved on Google, Microsoft or Meta's own login screen. We receive a limited access token, which you can remove from your account settings.
Your data stays in your accounts
Messages in your inbox, bookings in your calendar. The AI reads what it needs for each message and writes the result back to your system.
Nothing is used for training
We use commercial access to the models, which is not used for training.
Nothing is made up
If the answer is not in your information, the customer is told a person will reply, and the thread comes to you with a suggested answer.
It says it is AI
If anyone asks, it says it is an automated assistant and hands over to a person on request.
It is yours, even if you leave
It runs on your accounts, so nothing switches off if we stop working together.
Before it goes live
- We test it on your real messages and show you the answers side by side
- You get a written list of every service that touches your data
- For the first two weeks you approve every reply
- Every step is logged
The law
We work within the EU GDPR and UK GDPR. Your business remains responsible for its customers' data and we act only on your instructions. We do not build automated marketing calls or messages to strangers.
Questions about your data?
Ask us before we start. We will explain exactly what each step does.