An AI tool call is not consent: building visitor-approved question forwarding

Description and use

Tasda implemented a preview‑and‑confirm flow where the model proposes a question but the visitor authorizes its forwarding, ensuring proper consent and limiting data exposure.

Problem

Ensuring that a visitor’s decision, not the AI model’s suggestion, authorizes forwarding a question to the representative’s creator.

Outcome

The implementation separates model suggestion from user-initiated write, enforces context validation, limits data exposure to the short question, and ensures idempotent confirmations.

Workflow

• Model proposes question
• Server creates signed preview
• Client displays question with confirmation control
• Visitor confirms via authenticated request
• Server validates preview, context, and eligibility
• Server stores the request

Evidence

The model can prepare a question, but the visitor decides whether to send it.

Source

The preview expires after one hour and is signed.

Source

Only the short question shown in the preview is stored for the creator.

Source