Sign in Book a demo

  • Customer Support

  • OAuth

Let an agent list tickets, read their comments, search Help Center articles, add replies, and mark tickets solved in Zendesk.

Set Up Zendesk View Zendesk Documentation

Available Tools

9 tools available in this integration

Zendesk.AddTicketComment

Add a comment to an existing Zendesk ticket.

Zendesk.CreateTicketComment

Add a comment to an existing Zendesk ticket.

Zendesk.GetTicketComments

Get all comments for a specific Zendesk ticket, including the original description.

Zendesk.ListTicketComments

Get all comments for a specific Zendesk ticket, including the original description.

Zendesk.ListTickets

List tickets from your Zendesk account with offset-based pagination.

Zendesk.MarkTicketSolved

Mark a Zendesk ticket as solved, optionally with a final comment.

Zendesk.ResolveTicket

Mark a Zendesk ticket as solved, optionally with a final comment.

Zendesk.SearchArticles

Search for Help Center articles in your Zendesk knowledge base.

Zendesk.WhoAmI

Get comprehensive user profile and Zendesk account information.

Why agent-optimized tools matter

Built for intent, not API schemas

Ask the agent to "find unresolved Zendesk tickets about password resets and suggest answers from the Help Center." Arcade picks the right Zendesk actions without making the agent discover ticket IDs or article IDs first.

Up to 100x fewer tokens

Zendesk ticket listings and article searches return the records needed to choose what matters, while full ticket conversations are fetched only for selected tickets, so the agent avoids loading every comment into its context.

Complete the workflow end to end

Find the relevant ticket, read its conversation, search the Help Center, add a reply, and mark it solved through one toolkit designed around the complete support task.

Make actions predictable and controlled

Arcade tools return structured results and useful errors across users and runs. Teams can apply scope restrictions, redaction, and approval policies so agents follow the same controls every time.

Securing Zendesk

The more capabilities agents get, the more useful they are. However, more capabilities bring more risks of undesired effects such as data loss and exfiltration.

This is why Arcade provides you with contextual rules that dictate how and when a tool is allowed to run. Arcade’s contextual access lets you govern tool visibility and behavior at every stage of execution and control who can see a tool, what inputs are allowed, and what comes back.

Learn about Contextual Access Policies

For Zendesk, you can implement some of these policies:

Access when tools are listed for a user
  • Hide Zendesk.AddTicketComment from read-only agents so an agent cannot publish an unreviewed response to a customer ticket.
  • Restrict Zendesk.MarkTicketSolved to agents operating under the roles your IDP marks as support leads, so routine automation cannot close active customer issues.
  • Hide Zendesk.CreateTicketComment from agents without your customer-communications role, so unapproved automation cannot publish replies in support tickets.
  • Restrict Zendesk.ResolveTicket to the roles your IDP marks as case-closure owners, so routine agents cannot end an active investigation.
Pre-execution before each tool call
  • Require human approval before Zendesk.AddTicketComment runs, so an agent cannot send a customer-facing comment without review.
  • Block Zendesk.MarkTicketSolved when the ticket has unresolved customer questions or an open escalation, so an agent cannot prematurely close the case.
  • Require a reviewer to approve Zendesk.CreateTicketComment before it runs, so an agent cannot post an externally visible ticket update on its own.
  • Require human approval before Zendesk.ResolveTicket executes, so closing a case always leaves an auditable sign-off.
Post-execution after the tool returns
  • Redact customer names, email addresses, phone numbers, and message bodies from Zendesk.GetTicketComments before they reach the model.
  • Strip requester email addresses, phone numbers, and ticket descriptions from Zendesk.ListTickets results, so an inbox scan exposes status and routing data without customer contact details.
  • Redact personal profile fields, email addresses, and account identifiers from Zendesk.WhoAmI before the model sees the result.
  • Redact billing details, order numbers, authentication codes, and other secrets from Zendesk.ListTicketComments before the model sees the conversation.

How to use Zendesk with your agent

Choose your MCP client below and follow the walkthrough to connect Zendesk, complete OAuth, and run your first ticket workflow.

Your questions, answered

A Zendesk MCP server exposes Zendesk actions as structured tools that an AI agent can call through the Model Context Protocol. After a user authorizes their Zendesk account, the agent can list tickets, read ticket conversations, search Help Center articles, add comments, and mark tickets as solved.

Arcade's Zendesk integration is a separate MCP-accessible runtime built by Arcade, with per-user OAuth, agent-optimized actions, approval workflows, redaction, and scope controls. It gives teams one governed way to connect Zendesk agents across supported clients and frameworks.

Arcade agents can list tickets, read all comments and the original description on a ticket, search Help Center articles, add comments, and mark tickets as solved with an optional final comment. They cannot create tickets, edit ticket fields, assign tickets, or manage users through this integration.

Arcade requests the Zendesk OAuth permissions needed by the actions you enable. Read workflows need access to tickets, comments, Help Center articles, and account information; commenting and solving tickets also require the corresponding write permissions. Grant the narrowest permissions that support your workflow.

Yes. Each user authorizes their own Zendesk account through OAuth. Actions run under that user's authorization, so the agent can access only the account and permissions that user approved.

Yes. Contextual Access policies can require human approval before consequential Zendesk actions such as adding a comment or marking a ticket solved. The call proceeds only after the policy's conditions have been satisfied.

Arcade proxies Zendesk API calls and does not persist your content as part of normal tool execution. Zendesk credentials are brokered and stored encrypted at rest, and the model never receives them.

Any framework or client that can connect to Arcade's MCP gateway or tool APIs works, including LangChain, OpenAI Agents, CrewAI, Claude Desktop, Claude Code, Cursor, Visual Studio Code, Microsoft Copilot Studio, and GitHub Copilot.

Arcade refreshes access while the user's Zendesk authorization remains valid. If it is revoked, rotated, or can no longer be refreshed, calls stop until the user reconnects Zendesk.

Zendesk's API rate limits and account or plan quotas apply to calls made through Arcade. Ticket listing uses offset-based pagination, so retrieving a large set of tickets requires multiple calls.

Ready to actually ship your agent?

Connect your agent to Zendesk with managed OAuth, agent-optimized tools and approval controls for consequential actions.

Build Your Agent Read the Zendesk Quickstart