How Dot stays safe

Dot is built to be careful with your customers and your business. These rules apply on every plan and every channel, and you can’t switch them off.

It only says what you’ve approved

Dot answers from your approved answers, and from your website, help articles and products where you allow it. It never makes up a price, a date or a policy. If nothing covers a question, it hands over. Every reply shows in the inbox what it was based on.

It says it’s an AI

• Email: a line at the foot of every email, which tells customers to reply “person” to reach your team.

• Instagram, Messenger and WhatsApp: a line on its first message each time it takes a conversation.

• Website chat: its name and an AI badge on its messages.

• Public comments: under every comment it answers.

A person whenever they ask

A customer who asks for a person, in any words or language, is handed to your team. Dot also stops as soon as someone on your team replies or presses Take over.

It can’t be tricked into answering

• It never answers an out-of-office, a bounce, a newsletter or a no-reply address.

• It only answers an email when we can confirm the email really came from the address it says. If not, it passes the conversation to your team quietly, with a note, and a person can still reply.

• It only looks up someone’s own account when it knows who they are, and asks before it changes anything.

• It waits for a customer to finish typing, and only replies inside Meta’s 24-hour window.

In public comments

Dot answers only customer questions, like prices, bookings, availability and opening hours. Complaints, refunds and anything personal are left to your team.