your data, your rules

The weird trust problem with AI

You want AI to help with your life. Manage your calendar, draft emails, keep your notes organized, coordinate with your team. But you also want to know exactly what it can see, what it can do, and what it can't.

That's the trust problem with personal AI. The more access you give, the more helpful it gets. But the more access you give, the more you need to trust it. Most AI tools ask for broad permissions upfront and give you almost no visibility into what they actually do with your data.

And people are paying attention. AI data privacy concerns jumped 56% in just one year, with 233 reported AI incidents in 2024 according to Stanford's AI Index Report. Meanwhile, 77% of AI leaders say data privacy is a top concern for their strategy. The worry is real, and it's growing.

Cloa handles this differently. Instead of asking for everything upfront, Cloa lets you control permissions at a granular level. Per service, per action, at any time.

Three levels of permission. You pick.

Every integration in Cloa has three permission levels:

Off. The integration is disconnected. Cloa can't touch the service at all. Nothing flows in or out.

Confirm. Cloa can read data from the service but has to check with you before doing anything. So it can read your emails but will ask before sending a reply. It can look at your calendar but will confirm before creating an event.

Full auto. Cloa can both read and act on your behalf without asking. It schedules meetings, sends messages, and updates notes on its own.

You set these independently for each service. Your calendar might be full auto because you trust Cloa to handle scheduling, while your email is on confirm because you want to review outgoing messages first. Your Notion might be full auto for reading but confirm-only for writing.

This means you never have to give Cloa more access than you're comfortable with.

What confirm mode feels like in real life

Here's how it works in practice:

You tell Cloa: "remind the team about the deadline on Friday"

Cloa drafts a Slack message: "Hey team, just a reminder that the project deadline is this Friday. Let me know if you need anything."

Before sending, Cloa shows you the draft and asks: "Ready to send this to #general on Slack?"

You approve, and the message goes out. Or you edit it. Or you say "not now" and it waits. The point is you stay in control of every action that reaches the outside world.

In full auto mode, Cloa would just send it directly. Your call.

You control what Cloa knows

Transparency isn't just about permissions. It's about actually knowing what your AI has learned about you and being able to change it.

You can ask Cloa what it remembers about any topic and tell it to update or forget specific details. If Cloa got something wrong (maybe it misheard a name or mixed up a date) you can correct it right in conversation. Cloa learns from the correction and gets better.

This kind of control is rare. Most AI tools work like black boxes. You put data in, and something happens inside that you can't influence. Cloa keeps you in the loop.

How we think about your data

Cloa runs on three data principles:

Your data stays yours. What's stored in Cloa belongs to you. It's not sold, shared, or used outside of helping you. We don't train AI models on your data.

You can delete whenever. Want Cloa to forget something? Just tell it. Deletion is real, not just hidden.

We collect the minimum. Cloa only stores what it needs to help you. We use basic product analytics to improve the app experience, but we never analyze or share your conversations, personal data, or memory contents.

A recent study found that only 55% of organizations even require clear data ownership terms when working with AI vendors. We think that number should be 100%. You should always know who owns your data (hint: it's you).

Change your mind anytime

Permissions aren't permanent. You can adjust them whenever from the app:

• Start with confirm mode for a new integration, then move to full auto once you're comfortable
• Set an integration to off while you're on vacation
• Disconnect a service completely with one tap

There's no penalty for changing permissions. No "are you sure?" nag screens. No waiting period. Your changes take effect right away.

Why this matters more than ever

As AI assistants get more capable, proactively managing your day and remembering your whole life, the trust question gets bigger. An AI that can send emails, schedule meetings, and manage your notes has real power. That power needs to come with real control.

The privacy-enhancing technology market hit over $3 billion in 2024 and is projected to grow past $12 billion by 2030. People are investing in privacy because they know it matters.

We don't think the answer is "trust everything" or "trust nothing." It's graduated trust. Start small, build confidence, and expand access as the relationship develops. That's exactly what our permission model is designed for.

Your AI should work for you. And you should always know exactly how. That's the future of personal AI.

Frequently asked questions

What permission levels does Cloa offer?

Cloa offers three permission levels for each integration: Off (no access), Confirm (Cloa can read but asks before acting), and Full Auto (Cloa reads and acts without asking). You can set different levels for different services and change them whenever you want.

Can Cloa send emails or messages without my approval?

Only if you set that integration to Full Auto. In Confirm mode, Cloa drafts the message and shows it to you for approval before sending. In Off mode, Cloa can't access the service at all. You control the level of autonomy for each integration separately.

Does Cloa sell or share my personal data?

No. Cloa doesn't sell, share, or use your personal data for anything other than giving you personalized AI help. Your data isn't used to train AI models and isn't shared with anyone.

Can I see what Cloa knows about me?

Yes. You can ask Cloa what it remembers about any topic, and it will tell you. You can also tell Cloa to update or forget specific details at any time. You stay in control of what your AI knows.

What happens if I turn off permissions for a service?

The service disconnects immediately. Cloa can no longer see any data from it. Workflows that depend on that service will pause until you turn permissions back on. Any data already collected stays in your account unless you choose to delete it.