Trust and control
An agent that acts on your behalf has to say how far.
Each person connects their own assistant. It reaches whatever that person can already reach. Every change is attributed to them, logged, and can be undone.
Who sees what
Access follows the org chart you already maintain.
There is no second permission system. A franchisee’s assistant reaches that territory. A franchisor’s assistant reaches every location they already run. Change someone’s role and their assistant follows on the very next request.
Three roles
Same product. Different reach.
Franchisor · every territory they already run
Regional · the sites they already manage
Franchisee · one territory, own assistant
Each person’s connection
Connected once. Removed any time.
Each person connects their own assistant, in their own browser. If they leave, you remove their access and their assistant stops that moment. Nobody shares a login, and there is nothing to pass around.
The trail
Attributed to the human, never to a nameless bot.
Threads sit with the work. Writes are logged as the connecting person. You can read the preview, approve or reject in the management UI, or let the agent apply when that person already could have.
Will never
Commitments, not caveats.
- Never share one login across a team. Each person connects their own assistant. If they leave, you remove theirs. Nobody inherits a shared secret.
- Never exceed the connecting person’s reach. The agent sees what that person can already see. A franchisee does not become a franchisor by connecting chat.
- Never invent a second permission system. Scope is the org chart you already maintain. A live re-check on every request means a demotion is immediate.
- Never act without an attributable human. There is no anonymous agent identity in the audit trail. The write is yours.
- Never publish a credential. What Fraink shares publicly about your brand is public facts only. The credential is shown once to the person it belongs to, and never appears anywhere else.