Avatar Consent and Disclosure Gate (Onboarding)

Avatar Consent and Disclosure Gate (Onboarding)

1. 01

   Inventory every human likeness the client wants to appear in generated video

   List named executives, paid talent, and any stock or archival footage containing identifiable faces. Record the source file and the date the client supplied it.

2. 02

   Collect written likeness authorization before any avatar is trained or cloned

   A signed release per person, scoped to the specific campaign, channel, and contract term. Verbal approval from a marketing manager does not cover a talent release.

3. 03

   Confirm whether the client's data will enter a shared model training pool

   Forrester analysts argue private deployments outperform public tools for B2B marketing precisely because shared model access erases differentiation. Document the answer per account so the client can see where their footage lands.

4. 04

   Set the disclosure standard for synthetic presenters and dubbed audio

   Decide with the client whether on-screen labels, verbal attribution, or platform-level AI flags are required. Write the rule into the creative brief so editors do not improvise per asset.

5. 05

   Map language and localization scope against the client's regulated markets

   A platform advertising 160+ languages does not mean every market accepts synthetic spokespeople for financial or health claims. Flag jurisdictions where a human presenter stays mandatory.

6. 06

   Log the approved avatar roster and voice clones in the client's asset register

   Include the model version, the training date, and the person who authorized it. When a talent contract ends, the register is what tells you which assets must be pulled.

7. 07

   Brief the delivery team on the boundary between templated avatars and custom creative

   Templated presenters scale volume; differentiation comes from custom scripting and brand-specific styling. State which projects are allowed to use stock avatars and which require bespoke treatment.