Setting Up and Training Brand Guardian in Marq

What You'll Learn

How to access Brand Guardian settings
Creating and configuring brand profiles
Understanding auto-populated vs manual fields
Training the AI with example projects
Using accept/reject feedback to improve accuracy
Managing multi-brand setups

Accessing Brand Guardian Settings

Brand Guardian is configured through the admin settings in Marq.

Steps to Access

Log in to Marq with administrator credentials
Navigate to Admin Settings
Find the AI Tools section
Select AI Brand Compliance

Note: The feature currently shows as "AI Brand Compliance" in settings, but is being updated to "Brand Guardian" across the platform. You may see both names during the transition.

Who Can Access

Brand Guardian settings are restricted to administrators who have:

Admin-level permissions in Marq
Access to AI Tools settings
Responsibility for brand management

Setting Up Brand Profiles

Brand profiles contain the guidelines that Brand Guardian uses to check projects.

Creating a Brand Profile

For organizations with a single brand:

Access Brand Guardian settings
Create your primary brand profile
Give it a clear name (e.g., "Acme Corporation Brand")
Begin configuring guidelines

Multi-Brand Setup

Brand Guardian supports multiple brands, each with its own unique guidelines.

When to Use Multi-Brand

Parent company with subsidiary brands
Multiple product lines with distinct branding
Geographic variations of the same brand
Main brand plus sub-brands

Creating Additional Brand Profiles

In Brand Guardian settings, add a new brand profile
Name it clearly to distinguish from other brands
Configure guidelines specific to this brand
Train each brand's AI separately

Important: Each brand profile has its own set of guidelines and trained AI. They operate independently.

Configuring Brand Guidelines

Brand profiles include both auto-populated fields and manual configuration fields.

Auto-Populated Fields

Marq automatically pulls certain brand information from your existing brand assets.

Brand Colors

Source: Populated from brand assets already in Marq
What's included: Primary, secondary, and accent colors defined in your brand kit
Action needed: Review to ensure all approved colors are present

Typography

Source: Populated from brand assets in Marq
What's included: Approved brand fonts and typography standards
Action needed: Verify the fonts match your brand guidelines

Manual Configuration Fields

Some guidelines must be manually entered as they're not automatically detected from brand assets.

Logo Usage Instructions

Define rules for proper logo use:

Clear space requirements around logos
Minimum size specifications
Approved logo variations (full logo, icon, wordmark)
When to use each variation
Prohibited logo modifications
Background color requirements

Example guidelines:

"Logo must have clear space equal to the height of the 'M' in the wordmark on all sides"
"Use reversed white logo on dark backgrounds, full color logo on light backgrounds"
"Minimum logo size: 1 inch wide for print, 150px wide for digital"

Usage Instructions and Custom Guidelines

Add brand-specific rules beyond colors, fonts, and logos:

Tone and voice guidelines
Imagery standards
Layout and composition rules
Prohibited elements or approaches
Industry-specific compliance requirements
Context-specific usage rules

Example guidelines:

"Always use sentence case for headlines, never all caps"
"Product images must always include the blue accent color border"
"Social media posts must include the hashtag #BrandName"

Brand Guide Parsing Tool (Internal)

Marq has an internal tool that can help parse existing brand guide documents into the Brand Guardian fields.

What it does: Upload a brand guide PDF and the tool extracts information into the appropriate Brand Guardian sections.

Availability: This is currently an internal resource used by the Marq team. If you need help parsing a complex brand guide, contact your account team.

Future: There are plans to make brand guide uploading available directly in the product.

Training Your AI Brand Assistant

Once basic guidelines are configured, training the AI improves its accuracy in identifying brand compliance issues.

Why Training is Essential

While Brand Guardian uses your configured guidelines, training helps it understand:

Nuanced interpretations of rules
Context-dependent exceptions
Industry-specific standards
Your organization's specific brand preferences
Edge cases and unusual scenarios

Without training, Brand Guardian may flag things incorrectly or miss subtle brand issues. Training significantly improves accuracy.

The Training Flow

Brand Guardian includes a "Train Your AI Brand Assistant" flow specifically for improving accuracy.

Step 1: Select Example Projects

Choose projects that represent good brand compliance:

Projects that exemplify proper brand use
Various content types (presentations, brochures, social posts, etc.)
Different contexts (internal, external, promotional, informational)
Examples from different teams or use cases

Tip: Include 5-10 diverse examples for initial training.

Step 2: Upload to Training Flow

Access the "Train Your AI Brand Assistant" flow in Brand Guardian settings
Upload one of your example projects
Brand Guardian analyzes the project and generates compliance feedback

Step 3: Review AI Suggestions

Brand Guardian will show you callouts identifying what it thinks are brand compliance issues in your example project.

Within the training modal, you'll see:

Specific elements flagged
Reasoning for each flag
Options to accept or reject each suggestion

Step 4: Accept or Reject with Feedback

For each suggestion Brand Guardian makes, you can:

Accept

When to use: Brand Guardian correctly identified a brand issue

What happens: Confirms the AI's understanding is correct. This strengthens its ability to identify similar issues in the future.

Reject with Feedback

When to use: Brand Guardian incorrectly flagged something that's actually compliant

What happens: Your feedback teaches the AI why its suggestion was wrong. This prevents similar incorrect flags in the future.

Critical: Always provide feedback when rejecting. Simply rejecting without explanation doesn't teach the AI anything.

Providing Effective Feedback

The quality of your feedback directly impacts training effectiveness.

Good Feedback Examples

Scenario: AI flags a color as off-brand

Good feedback: "This teal (#17A2B8) is actually our approved accent color for digital content only. It's not in print guidelines but is correct for web and social."
Poor feedback: "This is fine" (doesn't explain why)

Scenario: AI questions logo placement

Good feedback: "Reversed white logo is correct for use on dark backgrounds per our brand guidelines section 3.2"
Poor feedback: "Logo is correct" (doesn't provide context)

Scenario: AI flags typography

Good feedback: "Social media graphics are allowed to use tighter line spacing than print materials. This 1.2x spacing is within our social media brand standards."
Poor feedback: "Wrong" (doesn't explain the exception)

What Makes Feedback Effective

Include these elements in feedback:

Why it's correct or incorrect: Explain the reasoning
Context or exceptions: Note when rules vary by medium or use case
Reference guidelines: Mention specific sections if applicable
Clarify intent: Help AI understand the principle behind the rule

How Accepting Updates Guidelines

When you accept Brand Guardian's suggestions during training, it updates your brand guidelines automatically.

Example: If Brand Guardian identifies that a project uses an unapproved color and you accept that suggestion, the system learns that this color should be flagged in future checks.

This creates a feedback loop where your brand guidelines become more comprehensive through the training process.

Ongoing Training and Refinement

Training isn't a one-time activity. Brand Guardian improves through continuous feedback.

When to Conduct Additional Training

Initially: Train with 5-10 example projects when first setting up
After brand changes: When guidelines are updated
When accuracy issues arise: If users report incorrect flags
For new content types: When expanding into new formats or channels
Periodically: Quarterly reviews help maintain accuracy

Learning from User Feedback

As users run brand checks on their projects, they can provide feedback on Brand Guardian's suggestions. This real-world feedback helps the AI improve.

Note: An "apply feedback" feature is being added to make it easier for users to provide feedback directly when they receive brand check results.

Monitoring Brand Guardian Performance

Regularly review how Brand Guardian is performing:

Are users finding the suggestions helpful?
Are there recurring false positives?
Are there brand issues being missed?
Do certain project types need more training?

Multi-Brand Training

When managing multiple brands, each brand profile needs separate training.

Training Each Brand Independently

Upload brand-specific example projects for each brand
Provide feedback specific to each brand's guidelines
Train for each brand's unique contexts and use cases
Monitor each brand's AI accuracy separately

Preventing Cross-Brand Confusion

Training each brand separately ensures Brand Guardian doesn't apply one brand's rules to another brand's projects.

Example: Brand A allows vibrant colors while Brand B requires conservative palettes. Training each separately ensures appropriate guidance for each brand's projects.

Testing and Rollout

Testing Brand Guardian

Before rolling out to all users:

Run checks on test projects: Use various project types to test coverage
Review accuracy: Verify suggestions are appropriate and helpful
Gather initial feedback: Have a small group test and provide input
Refine training: Address any issues identified during testing
Document common scenarios: Create guidance for users on interpreting results

Rolling Out to Users

When you're ready to enable Brand Guardian organization-wide:

Communicate the launch: Explain what Brand Guardian is and why it helps
Provide training: Show users how to run brand checks and interpret results
Set expectations: Clarify that Brand Guardian is advisory, not restrictive
Establish support: Make yourself available for questions initially
Gather feedback: Collect user input to continue improving

Best Practices

For Initial Setup

Review auto-populated fields carefully - they may need adjustments
Be comprehensive with manual guidelines - the more detail, the better
Use clear, specific language in guidelines
Include examples in usage instructions when possible

For Training

Use diverse example projects covering various content types
Always provide detailed feedback when rejecting suggestions
Train with both correct examples and known issues
Be consistent in your feedback across similar scenarios

For Ongoing Management

Review Brand Guardian accuracy quarterly
Update guidelines when brand standards change
Conduct additional training when expanding to new content types
Respond to user feedback about incorrect suggestions

Frequently Asked Questions

How long does initial setup take?

Basic configuration takes 1-2 hours. Training can take another 2-3 hours depending on how many examples you use and how detailed your feedback is.

Do I need to train again if brand guidelines change?

Yes. Update the guidelines in Brand Guardian settings and conduct training with examples showing the new standards.

What if Brand Guardian keeps flagging something incorrectly?

Provide detailed feedback each time explaining why it's correct. If issues persist, conduct additional focused training on that specific scenario.

Can I see what feedback has been provided during training?

Training feedback is used to improve the AI but isn't separately documented. Keep your own records if you need to track training history.

How many example projects should I use for training?

Start with 5-10 diverse examples. You can always add more training later as needed.

Does training one brand affect other brand profiles?

No. Each brand profile's AI is trained independently. Training for Brand A doesn't impact Brand B.

What if I don't have brand assets set up in Marq yet?

Set up your brand kit in Marq first. Brand Guardian pulls colors and typography from those brand assets.

Next Steps

Start Using Brand Guardian

Now that Brand Guardian is set up and trained, users can start running brand checks on their projects. The more it's used, the more feedback it receives, and the more accurate it becomes.

Learn More About Marqet Integration

Brand Guardian works seamlessly with Marqet automations:

Marqet Overview - Content automation platform
AI Forms - Check form-generated content for brand compliance
Playbooks - Automated content with brand checking

Need Help?

If you need assistance setting up Brand Guardian, want help with the brand guide parsing tool, or have questions about training, contact Marq support or your account team.

Visit our Help Center for more articles and resources.