---
title: "How agencies get client approval on AI-made brand work | Rainbrand"
description: "Agree the ground rules before the first draft, check the work against the client’s brand before they see it, tell them how it was made and what they’re deciding, and put each correction into their brand once."
url: "https://www.rainbrand.com/blog/client-approval-for-ai-brand-work/"
---

By [Tomas Ward](https://www.linkedin.com/in/tomasward1/)

Published 6 Nov 2026

5 min read

# How agencies get client approval on AI-made brand work

Agree the ground rules before the first draft, check the work against the client’s brand before they see it, tell them how it was made and what they’re deciding, and put each correction into their brand once.

## TL;DR

Agree AI use, disclosure and sign-off before drafting. Check the work internally, send the exact file with the decision you need, and retain the approval against that export. Put lasting brand corrections into the client’s maintained guidance; keep campaign-specific changes with the job.

## What AI changes about approval

AI-made variants need checks for redrawn logos, altered products and edits that change previously approved details. [OpenAI warns](https://developers.openai.com/api/docs/guides/image-prompting) that repeated edits can change details you intended to preserve.

Choose a few distinct directions for the client, explain AI use as agreed, and make reusable corrections reach the brand rather than only one draft.

## Agree the rules before the first draft

Record these in the statement of work or kickoff:

- Permitted tools and client material.
- What you will disclose about generated and human-made parts.
- Who decides public AI labels and adds them.
- Approver, included rounds and what counts as a yes on an exact file.
- Ownership and excluded uses, such as likenesses or unreleased products.

In the US, the [Copyright Office](https://www.copyright.gov/newsnet/2025/1060.html) says prompts alone do not establish copyright; perceptible human authorship, arrangements or modifications can qualify. If protection matters, decide what people make and obtain advice for the client’s markets.

[Client separation](https://www.rainbrand.com/blog/agencies-keep-client-brands-separate-in-ai-tools/) covers account access; the [onboarding checklist](https://www.rainbrand.com/blog/onboard-a-client-brand-into-ai-workflows/) covers unresolved brand guidance.

## Check the work before the client does

| Check against the brand | Confirm with the account’s sources/people |
| --- | --- |
| Correct client and current guidance | Product facts, prices and claim evidence |
| Original logo, palette and type | Rights to references, fonts and photography |
| Exact headline, call to action and required lines | No invented feature, person or event presented as real |

[Rainbrand Review](https://docs.rainbrand.com/guides/request-review) returns brand findings your agent can use in revisions. For a graphic, submit the final image, expected copy and constraints. Keep rights, product evidence and publishing approval alongside those findings.

Ask your agent

```
Request a Rainbrand Review of [the final image] against [the client brand], checking colour, typography, logo and copy. Use the exact text and constraints attached to this task. Revise the image from the findings and review it again, up to two rounds. Stop when the checks pass or a finding needs my judgment. Show the final image and anything still open.
```

[Image checking](https://www.rainbrand.com/blog/check-ai-images-against-brand-guidelines/) covers the detailed routine. An uncertain finding goes to someone who can resolve it.

## Send something the client can approve

Send the full-size file in its placement with:

- **Stage:** direction choice, draft changes or final approval.
- **Checks completed:** what is settled and what remains uncertain.
- **How it was made:** a factual account of generated, supplied and human-made parts.
- **Decision:** exactly what you need, from whom and by when.

| Round | File shown | Decision |
| --- | --- | --- |
| Direction | Two or three rough routes | Choose one |
| Draft | Chosen route at size/in placement | Required changes |
| Final | Exact export and labels | Approve or reject |

Keep the request pending until the agreed approver responds. Link the reply to that export; approval of an older draft does not cover a changed file.

Between rounds, edit the requested area and preserve approved assets. [OpenAI recommends](https://developers.openai.com/api/docs/guides/image-prompting) specifying what stays unchanged and compositing when a region must remain pixel-identical. Recheck the whole piece.

## Disclosing AI use

Tell the client as agreed. [Gain’s guide](https://blog.gainapp.com/ai-content-process/) distinguishes substantive AI-generated drafts from human work assisted with research, outlines or grammar.

Public labels depend on the market and use. The following reflects the sources checked in October 2026:

- **EU deepfakes:** [Article 50](https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-50) applies from [2 August 2026](https://ai-act-service-desk.ec.europa.eu/en/ai-act/timeline/timeline-implementation-eu-ai-act). Deployers—those using AI beyond personal, non-professional purposes—must disclose image, audio or video resembling [existing people, objects, places, entities or events](https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-3) that would falsely appear authentic. Evidently creative work still needs an appropriate disclosure that does not hamper its display or enjoyment.
- **EU public-interest text:** AI text intended to inform the public on matters of public interest needs disclosure unless it undergoes human review or editorial control and a person or organisation holds editorial responsibility. Agree whether agency or client supplies the label.
- **UK:** [ASA guidance](https://www.asa.org.uk/news/disclosure-of-ai-in-advertising-striking-the-balance-between-creativity-and-responsibility.html) gives no blanket AI-label requirement; ask whether omission would mislead. A label does not rescue a false product effect. [Advertisers remain responsible](https://www.asa.org.uk/news/ai-and-deepfakes-four-things-advertisers-need-to-know-before-they-hit-run.html).
- **US:** [FTC rules](https://www.ftc.gov/news-events/news/press-releases/2024/08/federal-trade-commission-announces-final-rule-banning-fake-reviews-testimonials) cover fake reviews/testimonials, including AI-made ones; [agencies can face responsibility](https://www.ftc.gov/business-guidance/resources/ftcs-endorsement-guides-what-people-are-asking).
- **New York:** Since June 2026, [commercial-ad creators must conspicuously disclose synthetic performers](https://www.nysenate.gov/legislation/laws/GBS/396-B) when they know they are present. These are digitally created or altered assets made to appear as human performers, without depicting a recognisable real person. Exemptions cover audio ads, AI used solely to translate a human performer’s language, and promotions for films, games and similar expressive works when the performer is used consistently with the work. Outlets that only carry the ad are excluded.
- **Platforms:** [Meta’s political/social-issue rules](https://transparency.meta.com/policies/ad-standards/SIEP-advertising/SIEP) require disclosure of specified realistic synthetic media. Check each placement’s rules. The [IAB framework](https://www.iab.com/news/updates-industry-framework-ai-transparency-disclosure-advertising/) is nonbinding and recommends targeted rather than universal disclosure.

For regulated categories or several markets, have the client’s counsel confirm. Record the decision and put any label on the final file submitted for approval.

## Fix each correction once

| Feedback | Where it belongs |
| --- | --- |
| Lasting logo, name, palette or asset correction | Client’s brand |
| This headline, crop, offer or date | Campaign brief/file |
| Preference on one piece | Job notes; confirm whether repetition makes it a rule |

Turn lasting feedback into a specific instruction and confirm it with the client. Rainbrand maintains that guidance across subsequent work: [prepare an edit](https://docs.rainbrand.com/guides/edit-brand), inspect it and apply or discard it. The next [Brief](https://docs.rainbrand.com/guides/request-brief) uses applied changes.

Remove retired files from the [Brand Library](https://docs.rainbrand.com/guides/manage-brand-library); add approved finals only when they represent the brand. External settings remain separate copies.

Ask your agent

```
Prepare a brand edit in Rainbrand for [the client brand] from the attached confirmed feedback. Keep task-specific choices with this job and preserve unrelated guidance. Show the proposed changes before applying them. Then list the affected Library files so I can confirm which retired ones to remove.
```

## Keep a record of what was approved

Keep the export, approver, reply/date, tools/references, brand guidance, disclosure decision and applied corrections together.

Save the [Rainbrand Brief](https://docs.rainbrand.com/api-reference/briefs/get-a-brief) with the file. Client sign-off stays in your email thread or proofing tool; Rainbrand does not collect it.

For the next recurring client job, agree the ground rules, [connect Rainbrand](https://docs.rainbrand.com/mcp/getting-started), check the draft before client review and turn lasting feedback into a brand edit. [Managing many client brands](https://www.rainbrand.com/blog/manage-many-client-brands-with-ai/) covers the wider roster.

## More in Agencies

- 6 Nov 2026 5 min read

    ### [How agencies keep client brands separate across AI tools](https://www.rainbrand.com/blog/agencies-keep-client-brands-separate-in-ai-tools/)

    Client brands get mixed in two ways: an agent applies the wrong client’s guidance, or one client’s material turns up in chats and projects where it doesn’t belong. Here’s how agencies guard against both in the AI tools they already use.

- 28 Oct 2026 5 min read

    ### [How agencies manage many client brands with AI agents](https://www.rainbrand.com/blog/manage-many-client-brands-with-ai/)

    Give each client job an explicit brand, current guidance and a handoff the next person can use. Keep brand changes separate from campaign feedback without replacing your delivery tools.

- 28 Oct 2026 4 min read

    ### [How to onboard a client brand into AI: an agency checklist](https://www.rainbrand.com/blog/onboard-a-client-brand-into-ai-workflows/)

    Collect the client's current sources, resolve conflicting guidance, create reusable brand context and check the first task. Keep the files, findings and unanswered questions in a handoff your team can use.

One brand, every agent.

[Connect your agents](https://docs.rainbrand.com/mcp/getting-started)