Security and trust

If an AI is going to work inside your world, you should know exactly what it can see and do.

Most executives already use AI in some form. Few have asked the harder question: what happens to my data once it enters these systems? Every BoyTwigg build starts from that question.

The architecture

Every build is its own environment.

Your entourage runs in an environment created for you alone. No shared infrastructure, no cross-client exposure, no third parties added without strong justification.

Data protection
Encrypted in transit and at rest. Your data is never shared, sold, or used to train AI models. It exists solely to power the work your agents do for you.
Access and credentials
Agents connect through scoped authorization you grant directly, never shared passwords. Credentials transfer through one-time secret links destroyed after use, and longer engagements include proactive key rotation.
Boundaries you define
Board communications, legal correspondence, personal financial accounts, or anything else you designate stays off-limits. These are rules your agents follow, not suggestions.
Your rights

Full control, at all times.

Revoke access instantly
Every connection can be disconnected from your own account settings at any time. Disconnection is immediate, with no waiting period and no residual access.
Access, export, or delete your data
Request a copy, take it in a portable format, or have it removed entirely.
No lock-in
If you discontinue, all agent access is terminated and your data is removed. No penalty, no data hostage.
How this compares

Consumer AI tools ask for trust. This is built to earn it.

Consumer AI tools
General-purpose. You paste information in and get a response, and depending on the platform, that information may improve future models. It is handing a stranger a document and asking for advice.
Your entourage
A controlled environment where agents operate inside your tools with defined permissions, defined boundaries, encrypted credentials, and no model training on your data. A trusted team member working within your systems, under rules you set.
Security FAQ

The questions executives ask before granting access.

Can my agent read all of my email?+

Only within the scope of permissions you authorize, and under explicit instructions defining what it engages with and ignores. An agent built for scheduling and follow-ups processes those threads. It does not browse your inbox. It follows its instructions and does not freelance.

Can it send emails or act without my knowledge?+
Is my data used to train AI models?+
Who else can see my data?+
What happens if there is a security incident?+
Can I shut everything down immediately?+
Insured operations

Backed by insurance written for AI work.

BoyTwigg AI carries dedicated coverage for AI operations and technical work: technology professional liability, cyber liability, and commercial general liability. The architecture protects your data, the process protects your systems, and the coverage stands behind both.

Tech professional liabilityCyber liabilityGeneral liability

Insurance documentation available upon request.

Corgi Insurance
Insured by Corgi →

Questions about security or data handling?

Ask them directly on the call. Real answers from the person who builds the systems.