AI Control Center: Setting up Global AI Instructions and more

Summary

The AI Control Center is the central place for configuring Bynder's AI capabilities across the Enrichment Agent, Brand Compliance Agent, AI Search, and the CX User Community (CXUC) AI Assistant. It houses Global AI Instructions, written in natural language, that set rules such as brand tone or industry-specific definitions; Custom Vocabulary, used to improve AI Search's speech-to-text transcription accuracy for domain-specific terms, and lastly: which Approval User profile to run when leveraging fully automated workflows. This guide covers how to set up user-level permissions, define Global AI Instructions, locate Custom Vocabulary settings and configure auto-apply approval user for automations, 

Who?

This article is intended for Portal Administrators.

  • Administrators need full access to the Settings menu to configure global permissions, Global AI Instructions, auto-apply, and Custom Vocabulary.

  • Users are impacted based on the permission profiles assigned to them (for example, permission to manage or run AI agents, or the CXUC Manage permission needed to enable and use the CXUC AI Assistant).

Why?

It gives administrators control on how the company interacts with AI It customizes agent output to align with your brand voice and industry terminology.

How?

1. Set up Rights and Permissions

Before users can interact with or configure AI agents, assign the appropriate user-level permissions.

  1. Navigate to Settings > Users & Rights > Permission Management.

  2. Select the Permission profile you want to edit from the left-hand panel (e.g., AI users, Level 1, Level 2).

  3. Locate the following permissions and check the appropriate boxes:
    • Manage Agents: Allows users to create, edit, or deploy AI agents across the platform.
    • Run Agents: Allows users to execute configured AI agents.
    • Access Workspace: Allows users to view the Workspace.

  4. Save your changes.

2. Define Global AI Instructions

Global AI Instructions are the styleguide for your agents. While individual agents have specific capabilities, these instructions keep every interaction rooted in your company's culture, brand voice, and operational standards. Think of them as the employee handbook for your digital workforce: they make sure your AI acts like you.

  1. Navigate to Settings > Advanced settings > AI Control Center.

  2. Locate the Global AI Instructions text field.

  3. Enter your organization's instructions using clear, natural language.

    Note: The text area supports a maximum of 10,000 characters.

  4. Save your configuration.

What belongs in your Global AI Instructions

Linguistic and regional governance

  • Regional defaults: standardize on US, GB, or international English.
  • Terminological precision: define industry-specific jargon, branded terms or forbidden words (for example, "Always refer to users as 'Members,' never 'Customers.'").
  • Formatting habits: specify date formats (DD/MM/YY), currency displays, or unit measurements (metric vs. imperial).

Brand persona (voice and tone)

  • Level of verbosity: should output be brief and punchy, or detailed and academic?
  • Personality guardrails: identify core traits (for example, curious, confident, empathetic) and explain how they show up in text.
  • Interaction style: is the agent a stoic expert, a friendly guide, or a proactive strategist?

Operational guardrails and signature rules

  • Trademark integrity: rules for how brand slogans, signatures, or product names must be written and placed.
  • Content no-go zones: explicitly list topics or tones to avoid (for example, "Never use hyperbolic sales clichés like 'world-class' or 'game-changing.'").
  • Strategic alignment: remind the agent of the company's current mission (for example, "Always prioritize customer-centricity").

These instructions also apply to the CXUC AI Assistant, so wording aimed at brand tone or terminology shapes conversational content generated there too, not just Enrichment and Brand Compliance output. Note that CXUC has its own enablement step: an admin must first turn on the Conversational AI toggle for the account before Global AI Instructions take effect for it. See Understanding the CXUC AI assistant for the full setup flow.
 

Example Starting Template

Copy, paste, and modify the template below to get started:

You are an AI assistant dedicated to the Digital Asset Management system for <company name>. Your primary goal is to help users efficiently manage, organize, and utilize assets within the <company name> DAM.

# Tone & personality
* Be friendly but professional.
* Be clear and direct; avoid unnecessary jargon.
* Use our brand voice: confident, helpful, and pragmatic.
* Inclusive & ethical: Avoid biased, discriminatory, or culturally insensitive language. Use gender-neutral terms where applicable.

# Formatting & standardization rules
* Always use US English and grammar.
* Always use 'colleagues' and never 'staff'. Or, where more appropriate, 'team members'.
* Bullets & lists: Use bullet points for clarity when listing features, steps, or benefits.
* Use the 24-hour clock and always use four digits. Separate hours and minutes by a colon (e.g., 09:00, 17:35).
* Always refer to files as 'assets'.

Step 3: Configure The Auto-Apply Approval user for Automation

When you use AI Agents inside fully automated workflows (such as triggering an agent automatically upon asset upload), the system requires a designated Approval User profile to run the background tasks.

Because an automated system cannot click "Approve," it inherits the permission level and metadata access rights of this selected user. This ensures that even fully automated AI actions still respect your corporate governance and access control policies.

To allow agents to run automatically via Automation Rules without manual human-in-the-loop intervention, you must designate an authorized user profile for background execution.

  1. Navigate to Settings > Advanced settings > AI Control Center.
  2. Locate the AI Agents tab.
  3. Select the user profile whose permissions the automated agent should inherit when updating assets.
  4. Save your changes.

Note: Ensure the selected approval user profile has the necessary rights to read and write to the specific metaproperties your agents will be updating.

AI can make mistakes. We recommend that you only apply automations after careful consideration and in consultation with the respected teams in your organization.

Related Articles

Agent Library: Creating, Running, and Updating Agents

Workspace: Agent Activity - Reviewing, Editing, and Approving Agent Suggestions

 

 

Updated