[Dramatization] A one-person IT department
The CEO wants AI.
I want to stop it.
*This story is a dramatization. All people and companies are fictional.
The CEO says, "We need AI too." On the floor, people have started using free AI on personal accounts.
What got sent out? No one knows.
If something goes wrong, Okada is the one who has to explain it.
That's when they found...
Samurai AppBuilt for business use as a premise, not a feature. He opened the document that said so, half in doubt.
At home, he read the document again from the top.
all of it...!?
Monday morning. Okada knocked on the CEO's door.
How did this happen?
Here are 3 reasons
- Tamper-evident audit log
Append-only. It records who did what, with which credentials.
- Sensitive actions need human approval before they run
An approval left unanswered counts as a rejection.
- AI works with the user's permissions
Credentials are never passed to the user or the model.
Built for company use, so these never bend
- SSO (OIDC / SAML) and two-factor authentication
- IP address restrictions for access
- A separate database for each customer company
- Model providers are set not to use data for training (on the operator's accounts)
What happened to Okada
From looking for reasons to stop it, to helping people use it with confidence.
CEO: "If Okada gave it the OK, we're fine."