How Otto works — and the promises we keep.
Otto talks to your customers. That’s a serious responsibility, so we made our rules public and plain. Four promises, no fine print.
Otto always discloses it’s an AI
On the first message of every chat, text, and voice call, Otto states clearly that it’s an AI assistant — not a person. Your customers are never deceived, and neither are you.
We never train models on your customers’ data
Your leads, conversations, and customer records are yours. We use them to run your business inside Sovereign — never to train a shared AI model, and never sold to anyone.
Nothing happens without your approval
Otto drafts and proposes. Every message that reaches a customer and every charge that moves money waits behind the gold approval pill until you say go. You are always the human in the loop.
Clear limits, honestly stated
Otto won’t invent facts, prices, or reviews; won’t pressure your customers; and won’t claim to be something it isn’t. When it’s unsure, it asks you instead of guessing.
Disclosure isn’t just polite — it’s the law.
AI-disclosure rules are arriving fast. Sovereign was built to meet them from day one, so you’re never exposed.
Effective Aug 2, 2026: AI systems must tell people they’re interacting with AI. Otto already does, on every channel.
Already requires bots to disclose in commercial and sales conversations. Otto’s first-contact disclosure covers it.
This is a plain-language summary, not legal advice. See our Terms and Trust Center.
What Otto won’t do
Guardrails we built on purpose.
Invent prices, dates, statistics, or testimonials
Send a customer message before you approve it
Charge a card or move money on its own
Pretend to be a human being
Use high-pressure or dark-pattern tactics
Make promises about work it can’t verify
Go deeper
Every promise above has a page that shows the receipts.
