Skip to content

DOUBLE CREDITS on your first month, or your first 3 months on聽annual

Claim now
Kyndrify
All articlesHow To

How to Create an AI Brand Spokesperson

Create a governed AI brand spokesperson with verified consent, claim limits, approval, disclosure, asset tracking, and an exit plan.

By the Kyndrify teamUpdated September 28, 202610 min read
Share

How to Create an AI Brand Spokesperson

You want one face for your brand. Here is the short answer.

An AI brand spokesperson is a set role for a real person. That person gives consent. The team then uses approved face and voice files to make scripted clips within the tool's rules.

This guide shows you how to plan, build, and manage that role. It uses Kyndrify as an example tool. But the steps work for any careful process.

What is an AI brand spokesperson?

It is not just a face that can be used again. It is a brand role with clear rules.

A real person agrees to be the face and voice. The team records approved source files. The plan sets where and how those files may be used. It also covers exit, review, and labels for each channel.

The plan says what the role may say. It names each channel and each person who can approve a clip. It also records the agreed controls for the real person.

A professional adult sitting at a desk with a laptop in a bright, modern conference room

Step 1: Define the role and scope

Start with a clear plan. Write down what the spokesperson will do.

  • Role. Will they welcome new users? Explain products? Appear in ads?
  • Claim scope. What can they say? Only approved product facts? Personal stories?
  • Channels. List every place the content will go. Ads, social, help pages, email.
  • Duration. How long will you use this likeness? One campaign? One year?

A written scope gives the person and the production team the same reference for review.

Step 2: Choose the right person

Pick one person. They must fit the role you defined.

You have three common paths.

  1. Use your own face. This works for founders.
  2. Use a founder or leader. They already speak for the brand.
  3. Use a consenting team member. They must agree freely.

Choose someone whose natural style matches your brand tone. Calm, warm, or direct. Do not force a mismatch.

The person must give real consent. We cover that next.

Step 3: Get written consent and set usage scope

Do not treat consent as a quick yes. Record the terms in writing.

Work with a qualified advisor. This article does not state the law. Topics to review include:

  • What you will record (face, voice, or both).
  • Where the content will appear.
  • How long you can use it.
  • What happens if the person leaves or revokes consent.
  • How you will store and delete the source files.

Store the signed record with access controls. Check it before each new campaign.

Step 4: Plan the exit and revocation

People leave jobs. Plans change. Write the exit steps now.

Ask qualified counsel what the agreement should cover in your region and situation. Topics to discuss include:

  • How the person can revoke consent.
  • How fast you must stop using their likeness.
  • Who owns the existing content.
  • Whether you can keep old videos live.

Record the approved exit process before the first campaign, then connect it to the asset register and publishing owners.

A couple working together to sign important legal documents at a desk in an office setting

Step 5: Capture the source material

Now you record the person. You need two things.

  1. A photo or headshot.
  2. A short voice clip.

Kyndrify uses these to build a Twin. Each new video is called a Render. A Render returns a video download and a hosted link. Longer scripts can be split and stitched. The documented setup is on the Kyndrify talking-head video page.

Follow the current capture instructions in the product and reject source material that does not pass the setup check.

Step 6: Build the brand voice guide

The role also needs a short voice guide.

Write a short brand voice guide. Include:

  • Tone. Friendly, formal, or playful.
  • Pace. Fast and energetic? Slow and calm?
  • Word choices. Words you always use. Words you never use.
  • Sentence length. Short and plain? Or longer and more detailed?

Give the same guide to each writer and reviewer. Use it as part of the clip test.

Step 7: Set the script source of truth

Each Render starts with a script. Keep one approved copy.

Store scripts in one place. Label each with:

  • The campaign name.
  • The date it was written.
  • The person who approved it.

Link the approved script to the Render. Do not reuse an old draft by mistake.

Step 8: Run an acceptance test

Before you publish, test the output.

Watch the Render. Ask:

  • Does the face move naturally for this script?
  • Does the voice match the brand guide?
  • Are the words clear and easy to follow?
  • Does the message match the approved script?

If something feels off, adjust the script or the settings. Test again. Only move forward when it passes.

Step 9: Build a disclosure matrix by channel

Use one check for AI-media labels. Use a second check for ad or endorsement rules. Each channel and region may differ.

Close-up of hands typing on a laptop with smartphone on desk

Here is a simple matrix to start. Check each platform for updates.

Channel Disclosure method Notes
TikTok Enable creator label before posting. TikTok may also auto-label content with attached Content Credentials. Source
Other social platforms Check each platform's current AI label rules. Rules change often. Recheck before each campaign.
Endorsement with a material connection (U.S.) Follow current FTC guidance for a clear, conspicuous disclosure with the endorsement. For a video endorsement, FTC guidance says the material-connection disclosure should be in the video, not only in the description. Source
Website embeds Use visible on-page labels. Add text near the video player.

Kyndrify is rolling out signed Content Credentials and invisible provenance, but they are not guaranteed on every file. Free outputs currently carry a visible Kyndrify mark, while paid outputs omit that mark. Treat these file-level signals as one part of the plan. They do not replace a channel label or an endorsement disclosure when one is required.

Step 10: Set the approval workflow

Every Render needs a human check. Build a simple workflow.

  1. Writer drafts the script.
  2. Brand lead approves the script.
  3. Render is created.
  4. Brand lead watches the Render.
  5. Legal or compliance checks it (if needed).
  6. Final approval for publish.

Assign the required approvers by claim risk and channel. A routine factual update may need a different path from a regulated claim or paid endorsement.

Step 11: Keep an asset register

Track every Render you make. A simple spreadsheet works.

Columns to include:

  • Render ID or file name.
  • Date created.
  • Script used.
  • Channels where it was published.
  • Disclosure labels applied.
  • Expiry or review date.

Use this register to locate published content during routine review, correction, expiry, or an approved revocation process.

Close-up of a person writing at an office desk with documents and coffee, symbolizing productivity

Step 12: Set a maintenance cadence

Your spokesperson is not a one-time project. It needs regular care.

  • Review the asset register on a cadence tied to how often facts and offers change.
  • Recheck platform disclosure rules before each new campaign or material channel change.
  • Review consent on the schedule set by the agreement and qualified advice.
  • On exit or revocation, follow the approved process and document each action.

Step 13: Plan for incidents

Things can go wrong. Have a simple response plan.

  • Wrong script used. Pull the content. Post a correction.
  • Consent revoked. Follow your exit plan. Remove content fast.
  • Platform flags content. Check the disclosure. Fix and re-upload if allowed.
  • Negative feedback. Review the content. Decide if it still fits your brand.

Write down who assesses, removes, corrects, approves, and records each response.

Spokesperson brief template

Use this template for every new campaign.

  • Campaign name:
  • Spokesperson (person's name):
  • Role in this campaign:
  • Approved claims:
  • Channels:
  • Disclosure method per channel:
  • Script source of truth link:
  • Consent agreement reference:
  • Review date:

Pre-render and pre-publish checklist

Before you render:

  • Script matches approved claims.
  • Script follows brand voice guide.
  • Consent agreement is current.
  • Channel list is confirmed.

Before you publish:

  • Render passed acceptance test.
  • Disclosure is added per channel matrix.
  • Approval workflow is complete.
  • Asset register is updated.

Pricing snapshot (September 2026)

Kyndrify plans use a shared credit balance, and pay-as-you-go credit packs are also available. Free provides 8,400 one-time signup credits that never expire, with no card required. Free renders use preset avatars and voices; own-face rendering requires an eligible plan such as Plus. This is a snapshot. Check the official page for current details. Kyndrify pricing page

Illustration: estimating your Render volume

This is not a promise. It is a simple way to think about volume.

Estimate monthly volume as approved scripts 脳 expected Render attempts per script, then add planned variants and maintenance rerenders. Replace every input with your own acceptance data and current Render price.

Frequently asked questions

What is an AI brand spokesperson?

It is a governed brand role. A real person gives consent. You record their face and voice. You use that recording to make new video clips from scripts you write.

Do I need to capture new source material for every script?

Kyndrify documents Twin setup from a photo and short voice clip, followed by scripted Renders. Recapture when the product requires it or when identity, consent, brand, or acceptance needs change.

Can I use any person's face?

Use a real person only within verified consent, written usage scope, vendor terms, platform rules, and applicable law. TikTok also prohibits specific AI likeness uses even when labeled. Get qualified advice for the campaign.

How do AI labels and endorsement disclosures differ?

An AI label explains synthetic media. An endorsement disclosure explains a material connection. Follow each channel's AI rules; when U.S. FTC endorsement guidance applies, make the material-connection disclosure clear and put a video disclosure in the video.

What should I check before choosing a tool?

Kyndrify recommends checking consent and likeness rights, voice-clone consent, disclosure and provenance features, pricing, and output ownership. Verify these with any vendor you consider. Kyndrify pricing page

Disclaimer

This article is not legal advice. Rules for consent, likeness rights, and disclosure vary by region. Work with a qualified advisor. Always verify vendor terms before you buy.

Related reading

ai brand spokespersonai spokespersonvirtual spokesperson

Make your first video without filming.

Say what you need and the studio makes it: video, images, voices. Start free, no credit聽card.