Zammad Integration for Open Ticket AI

Open Ticket AI (OTAI) includes a plugin for Zammad, allowing OTAI to read and update tickets through the Zammad REST API. This enables AI-driven routing, prioritization and categorization directly inside Zammad. The current implementation works, but it is not fully tested yet and may require adjustments in real-world installations.

Architecture

The integration follows the standard OTAI plugin pattern:

• a separate plugin package: otai_zammad
• a ZammadTicketsystemService (Injectable)
• a ZammadPlugin registering the service
• configuration in config.yml
• OTAI writes predictions back into Zammad through the service’s REST calls

This structure is identical to all OTAI ticket system plugins (e.g. OTOBO/Znuny, Freshdesk, OTRS).

Zammad Plugin Structure

otai_zammad/
src/
otai_zammad/
zammad_ticket_system_service.py
plugin.py
pyproject.toml

plugin.py

from open_ticket_ai import Injectable, Plugin

from otai_zammad.zammad_ticket_system_service import ZammadTicketsystemService


class ZammadPlugin(Plugin):
    def _get_all_injectables(self) -> list[type[Injectable]]:
        return [
            ZammadTicketsystemService,
        ]

The plugin exposes exactly one injectable: ZammadTicketsystemService.

ZammadTicketsystemService Parameters

The service accepts these fields:

base_url: The base URL of the Zammad instance (e.g. https://helpdesk.example.com)
access_token: Personal Access Token used for authentication
timeout: Optional HTTP timeout in seconds
verify: TLS verification flag or path to CA bundle

These map directly to the service’s params model.

Example Configuration

(You just change “use” to zammad:ZammadTicketsystemService)

ticket_systems:
  zammad:
    use: zammad:ZammadTicketsystemService
    params:
      base_url: https://your-zammad-domain
      access_token: "{{ get_env('ZAMMAD_TOKEN') }}"
      timeout: 10
      verify: true

After this, OTAI will automatically load the plugin via the entry point in your pyproject.toml:

[project.entry-points."otai.plugins"]
otai_zammad = "otai_zammad.plugin:ZammadPlugin"

How OTAI Uses the Zammad Service

1. OTAI fetches new or updated Zammad tickets
2. AI models classify queue, priority or custom labels
3. OTAI calls update_ticket(...) on the Zammad API
4. Zammad updates the ticket
5. Agents continue their normal Zammad workflow, now powered by OTAI

The process is identical to OTOBO/Znuny, Freshdesk, OTRS or any other OTAI ticket system plugin.

Current Status

The Zammad integration is implemented, but:

• it is not fully tested
• API coverage may be incomplete
• real-world Zammad setups may require additional adjustments
• performance characteristics (pagination, search endpoints, large result sets) still need benchmarking

You can already use the plugin for prototyping and internal testing, but production use should wait until further validation.

Benefits

• AI classification entirely on-premise
• no Zammad cloud extensions or external AI needed
• integrates cleanly into existing OTAI workflows
• same plugin architecture as all other OTAI services