Privacy
We designed Ask My Agent to protect your privacy — mainly from ourselves!
The problem
The information you give your personal agents can (obviously) be highly-personal, as can the conversations that people will have with them. You might not want that information in our hands.
The solution
You get what we'll call a “fenced-off personal install” of Ask My Agent, one that runs on your own personal accounts on global platforms that you already trust. You create these accounts, share access for us to set things up, and when we're done you un-share access and we are effectively outside the fence.
Step by step
How it works
- 1
You create personal accounts on the global platforms that the Ask My Agent system runs on.
This means Google Cloud (a part of Google most people don't use but that everyone with a Google account has access to) and other global platforms, such as one called Cloudflare. You won't need a credit card to create any of these accounts, and it'll take a minute or so for each.
- 2
You share access to these accounts to our team.
Typically these are newly-created accounts, meaning there's nothing in them yet (sensitive or otherwise), and sharing access simply means adding our email.
- 3
Our team sets everything up in the background
This happens while you are receiving training on the setup call. If we run into any issues, like the level of access you've granted being not quite right, you'll be on the call to help us address them.
- 4
We test to make sure it all works.
Goes without saying!
- 5
You un-share access
This is as simple as un-sharing a Google Doc; you remove our setup email and we can no longer access. When visitors start chatting with your agents, that data remains between you and the platforms that host your personal accounts.
- 6
Everything continues to work
We'll have set things up in such a way that the system will continue to work without us needing to retain access, and that we can update it for you without being able to access private data.
You hold the keys
Blind updates from our team
Later you'll have entered sensitive details about yourself, and perhaps your agents will have had sensitive conversations with visitors.
For Ask My Agent system updates, you can invite our team back in a way that means we cannot see this sensitive data (but can still update the system). The important detail here is that you booted us out earlier, and this makes it possible for you to invite us back on your terms.
1We install the workdesk
Our team sets the whole system up across your accounts, while you watch on the call.
2You change the lock on the top drawer
You un-share access. Everything sensitive that goes in from here — your instructions, your chats — is behind a lock only you hold.
3You invite us back to fix the desk
We can mend a wobbly leg without ever opening that drawer. Updates happen; the confidential part stays shut.
A fair question
Doesn't this just mean trusting big tech with my privacy instead of trusting a startup? 🤔
Largely, yes! But, realistically, you are most likely already trusting one or more of these platforms with your private data, and have been for years, perhaps decades.
This isn't to say that startups like ours can't be trusted to get data security right. Many startups are exceptional here. We ourselves work hard on security. But with data that is both highly-personal and also needs to be securely shared around to various friends and other human helpers, it can make good sense to rely on the personal-account infrastructure, developed over decades, of global tech platforms.