BoldDesk’s AI features—AI Search, AI Agent, and AI Copilot—deliver answers and suggestions that are strictly scoped to each user’s permissions. The system checks roles, brands, visibility, and knowledge source access before producing results. This ensures that agents, contacts, and groups only see what they’re allowed to see, keeping responses secure, relevant, and compliant.

Let’s look at how this works in each AI component:

Help Center (Knowledge Base)

AI Search (Generative)

When AI Search is enabled for the Help Center, search results and generated summaries respect:

• Article visibility (Public, Internal, or Restricted to roles/brands).
• Category/Section scoping you’ve set in the Knowledge Base module.
• Contact/Contact Group Knowledge Base‑access scope when Help Center sign‑in is required.

Explore further instructions about How to Enable AI Assistance in Knowledge Base Search.

Live Chat

AI Agent (Chat)

AI Agent answers questions 24/7 using the knowledge sources you approve. It always enforces BoldDesk permissions:

• Source selection: Choose permitted KB categories and file sources under AI Agent configuration.
• Brand isolation: AI Agent responds with content belonging to the active brand.
• Ticket/Contact context: If the user is signed in via the Help Center, the agent can use contact profile and ticket context (when enabled) for personalized answers without revealing restricted information.

Ways of User Identification in Live Chat

• If the live chat is embedded within the BoldDesk customer support portal, custom JavaScript can be added to automatically update the logged-in user details.
• If the live chat is implemented on an external site, the team must manually include custom scripts to add the user identity.

Explore further instructions about Live Chat User Verification with JWT in BoldDesk.

Agent Workspace

AI Copilot (Reply Assist)

AI Copilot assists agents inside tickets with suggested replies and context insertions. It adheres to:

• Agent role permissions (e.g., ability to view Internal/Restricted articles).
• Ticket brand & visibility rules.
• Field‑level access (only surfaces data the agent can normally see in the ticket UI).

Admin Settings (Security & Governance)

To keep AI responses aligned with your governance:

• Roles & Permissions: Limit who can view Internal KB, see Restricted articles, or access brand‑specific data (Admin → Roles & Permissions).
• Brands: Maintain isolation; AI operates in the active brand context (Admin → Brands).
• Help Center Access: Require sign‑in and map Contact/Contact Group visibility (Admin → Help Center → Access Control).
• Data Sources: Approve only vetted sources for AI (Admin → AI → Data Sources).
• Audit: Monitor AI usage alongside ticket history in Tickets → Activity/History and Help Center analytics.

Explore further instructions about Managing Roles and Permissions in BoldDesk.

FAQs

Q1: Can AI Search show an Internal or Restricted article to a public user?
No. AI Search follows Help Center visibility. If a user lacks access, the article (and its summary) will not be shown or cited.

Q2: Can I allow AI Agent to use only specific KB categories?
Yes. Under AI Center → AI Agent → Knowledge Sources, select the exact categories you want included.

Q3: Does AI Copilot override SLAs or ticket rules?
No. AI Copilot provides suggestions and context only. SLAs, statuses, and automations remain enforced by your configured policies.

Q4: Can I disable AI for a specific brand while keeping it on for others?
Yes. Configure AI per brand. Assign or withhold AI Agent and AI Search at the brand level.

Q5: Can the AI read draft KB articles?
No. Only published and included sources are used.

Related Articles

1. BoldDesk AI Copilot: Your Intelligent Assistant for Customer Service
2. Enabling AI Agent in Live Chat
3. How to Add Data Sources for AI in BoldDesk